See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/328100574
...
305 downloads
2246 Views
4MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/328100574
Fractional Order Processes: Simulation, Identification, and Control Book · September 2018 DOI: 10.1201/9780429504433
CITATIONS
READS
0
81
1 author: Seshu Damarla National Institute of Technology Rourkela 14 PUBLICATIONS 12 CITATIONS SEE PROFILE
All content following this page was uploaded by Seshu Damarla on 19 January 2019. The user has requested enhancement of the downloaded file.
Fractional Order Processes
This page intentionally left blank
Fractional Order Processes Simulation, Identification, and Control
by Seshu Kumar Damarla and Madhusree Kundu
MATLAB® is a trademark of The MathWorks, Inc. and is used with permission. The MathWorks does not warrant the accuracy of the text or exercises in this book. This book’s use or discussion of MATLAB® software or related products does not constitute endorsement or sponsorship by The MathWorks of a particular pedagogical approach or particular use of the MATLAB® software. CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742
© 2019 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group, an Informa business No claim to original U.S. Government works Printed on acid-free paper This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint. Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data Names: Damarla, Seshu Kumar., author. | Kundu, Madhusree, author. Title: Fractional order processes : simulation, identification, and control / Seshu Kumar. Damarla and Madhusree Kundu. Description: Boca Raton : Taylor & Francis, a CRC title, part of the Taylor & Francis imprint, a member of the Taylor & Francis Group, the academic division of T&F Informa, plc, 2018. | Includes bibliographical references and index. Identifiers: LCCN 2018021873| ISBN 9781138586741 (hardback : acid-free paper) | ISBN 9780429504433 (ebook) Subjects: LCSH: Fractional calculus. | Intelligent control systems–Mathematics. | Chaotic behavior in systems–Mathematical models. Classification: LCC QA314 .D295 2018 | DDC 515/.83–dc23 LC record available at https://lccn.loc.gov/2018021873
Typeset in Palatino by Integra Software Services Pvt. Ltd. Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the CRC Press Web site at http://www.crcpress.com
Dedicated to the fond memory of my beloved father, late Venugopalarao Damarla Seshu Kumar Damarla Dedicated to my students Madhusree Kundu
This page intentionally left blank
Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi About the Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxiii
1 Mathematical Postulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 Special Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1.1 Gamma Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1.2 Beta Function. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.1.3 Mittag-Leffler Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.1.4 Hypergeometric Function . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.1.5 Error Function and Complementary Error Function . . . . 4 1.1.6 Bessel Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.2 Definitions and Properties of Fractional-Order Operators . . . . . . 6 1.2.1 Grunwald-Letnikov (GL) Fractional-Order Derivative . . 6 1.2.2 Riemann-Liouville (RL) Fractional-Order Integral. . . . . . 7 1.2.3 Riemann-Liouville Fractional-Order Derivative. . . . . . . . 8 1.2.4 Caputo Fractional-Order Derivative . . . . . . . . . . . . . . . . . 8 1.2.5 Properties of GL, RL, and Caputo Fractional-Order Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 1.3 Laplace Transforms of Fractional-Order Operators . . . . . . . . . . 10 1.4 Fractional-Order Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 1.5 Fractional-Order PIλ, PDµ, and PIλDµ Controller . . . . . . . . . . . . . 14 1.6 Triangular Orthogonal Functions . . . . . . . . . . . . . . . . . . . . . . . . . 15 1.6.1 Review of Block Pulse Functions . . . . . . . . . . . . . . . . . . . 15 1.6.2 Complementary Pair of Triangular Orthogonal Function Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 1.6.3 Expansion of Two Variable Function via TFs . . . . . . . . . 20 1.6.4 The TF Estimate of the First-Order Integral of Function f ðtÞ . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 1.6.5 The TF Estimate of Riemann-Liouville Fractional-Order Integral of f ðtÞ . . . . . . . . . . . . . . . . . . . . 24 1.6.6 Error Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 1.6.7 MATLAB® Code for Generalized Triangular Function Operational Matrices . . . . . . . . . . . . . . . . . . . . . 29 1.7 Triangular Strip Operational Matrices for Classical and Fractional Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 vii
viii
Contents
1.7.1 1.7.2 1.7.3
Operational Matrix for Classical Derivative . . . . . . . . . . 31 Operational Matrix for Fractional-Order Derivative . . . 33 MATLAB Code for Triangular Strip Operational Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .34 2 Numerical Method for Simulation of Physical Processes Represented by Weakly Singular Fredholm, Volterra, and Volterra-Fredholm Integral Equations . . . . . . . . . . . . . . . . . . . . . 35 2.1 Existence and Uniqueness of Solution . . . . . . . . . . . . . . . . . . . . . 38 2.2 The Proposed Numerical Method . . . . . . . . . . . . . . . . . . . . . . . . . 41 2.3 Convergence Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 2.4 Numerical Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 2.4.1 Investigation of Validity and Accuracy . . . . . . . . . . . . . . 48 Example 2.1: The weakly singular (WS) Fredholm-Hammerstein integral equation (IE) of 2nd kind . . . . . . . . . . . . . . 48 Example 2.2: WS linear Fredholm IE of 2nd kind . . . . . 50 Example 2.3: WS Fredholm-Hammerstein IE of 1st kind . . . . . . . . . . . . . . . . . . . . . . . . 50 Example 2.4: WS Volterra-Fredholm-Hammerstein IE of 2nd kind. . . . . . . . . . . . . . . . . . . . . . . . 51 Example 2.5: WS Volterra-Hammerstein IE of 2nd kind . . . . . . . . . . . . . . . . . . . . . . . 51 2.4.2 Numerical Stability Analysis. . . . . . . . . . . . . . . . . . . . . . . 53 Example 2.6: WS linear Volterra-Fredholm IE of 2nd kind . . . . . . . . . . . . . . . . . . . . . . . . . 53 2.4.3 Application of Proposed Method to Physical Process Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Application 2.1: Heat radiation in a semi-infinite solid. . . . . . . . . . . . . . . . . . 54 Application 2.2: Hydrodynamics . . . . . . . . . . . . . . . . . . . 56 Application 2.3: Lighthill singular integral equation. . . . . . . . . . . . . . . . . . . . . . . . . . 58 2.5 MATLAB® Codes for Numerical Experiments . . . . . . . . . . . . . . 61 2.6 Summary of Deliverables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .70 3 Numerical Method for Simulation of Physical Processes Modeled by Abel’s Integral Equations . . . . . . . . . . . . . . . . . . . . . . . . . 73 3.1 Existence and Uniqueness of Solution . . . . . . . . . . . . . . . . . . . . . 76 3.2 The Proposed Numerical Method . . . . . . . . . . . . . . . . . . . . . . . . . 77 3.3 Convergence Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
Contents
ix
3.4
Numerical Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 3.4.1 Investigation of Validity and Accuracy . . . . . . . . . . . . . . 85 3.4.2 Numerical Stability Analysis. . . . . . . . . . . . . . . . . . . . . . . 87 3.4.3 Application to Physical Process Models Involving Abel’s Integral Equations. . . . . . . . . . . . . . . . . 89 Application 3.1: Cyclic voltammetry for the reversible deposition of metals on a solid planar macroelectrode . . . . . . . . . . . . . . . . . . . . 89 Application 3.2: Cyclic voltammetry for reversible charge transfer at a planar macroelectrode . . . . . . . . . . . . . . . . . . . . 90 Application 3.3: Potential step chronoamperometry for an irreversible charge transfer at a spherical electrode . . . . . . . . . . . . . 91 Application 3.4: Cyclic voltammetry for an irreversible charge transfer at a spherical electrode . . . . . . . . . . . . . 92 Application 3.5: Cyclic voltammetry for the catalytic mechanism at a planar electrode . . . . . 93 3.5 MATLAB® Codes for Numerical Experiments . . . . . . . . . . . . . . 96 3.6 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 4 Numerical Method for Simulation of Physical Processes Described by Fractional-Order Integro-Differential Equations . . . 109 4.1 Existence and Uniqueness of Solution . . . . . . . . . . . . . . . . . . . . 110 4.2 The Proposed Numerical Method . . . . . . . . . . . . . . . . . . . . . . . . 112 4.3 Convergence Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 4.4 Numerical Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 Case study 4.1: Fractional-order Fredholm-Hammerstein integro-differential equation . . . . . . . . . . . . . . . 122 Case study 4.2: Fractional order Volterra-Fredholm integro-differential equation . . . . . . . . . . . . . . . 122 Case study 4.3: Fractional-order population growth model. . . 123 Case study 4.4: Fractional-order integro-differential equations in anomalous diffusion process. . . . 126 4.5 MATLAB® Codes for Numerical Experiments . . . . . . . . . . . . . 128 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 5 Numerical Method for Simulation of Physical Processes Represented by Stiff and Nonstiff Fractional-Order Differential Equations, and Differential-Algebraic Equations. . . . 135 5.1 Existence and Uniqueness of Solution . . . . . . . . . . . . . . . . . . . . 136
x
Contents
5.2 5.3 5.4
The Proposed Numerical Method . . . . . . . . . . . . . . . . . . . . . . . . 138 Convergence Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 Numerical Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 5.4.1 Investigation of Validity and Accuracy . . . . . . . . . . . . . 140 Example 5.1: Simple linear multiorder Fractional differential equation (FDE). . . . . . . . . . . . 140 Example 5.2: Complex linear high-order FDE . . . . . . . 142 Example 5.3: Complex linear low-order FDE . . . . . . . . 143 Example 5.4: Nonlinear multiorder FDE . . . . . . . . . . . . 144 Example 5.5: Linear multiorder FDE with variable coefficients . . . . . . . . . . . . . . . . . . 144 Example 5.6: Linear fractional-order differential-algebraic equation (FDAEs) . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 Example 5.7: Nonlinear FDAEs . . . . . . . . . . . . . . . . . . . 146 Example 5.8: System of nonlinear FDEs. . . . . . . . . . . . . 147 5.4.2 Application to Physical Processes Described by FDEs and FDAEs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 Application 5.1: Bagley-Torvik equation . . . . . . . . . . . . 147 Application 5.2: Two-point Bagley-Torvik equation . . 148 Application 5.3: Plant-herbivore model . . . . . . . . . . . . . 148 Application 5.4: Financial mode . . . . . . . . . . . . . . . . . . . 151 Application 5.5: Epidemiological model for computer viruses . . . . . . . . . . . . . . . . . . . . . . . . . . 152 Application 5.6: Chemical Akzo Nobel problem . . . . . 156 Application 5.7: Robertson’s system . . . . . . . . . . . . . . . 160 Application 5.8: High Irradiance Responses (HIRES) of photo morphogenesis . . . . . . . . . . . 160 ® 5.5 MATLAB Codes for Numerical Experiments . . . . . . . . . . . . . 173 5.6 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
6 Numerical Method for Simulation of Fractional Diffusion-Wave Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191 6.1 The Proposed Numerical Method . . . . . . . . . . . . . . . . . . . . . . . . 192 6.2 Convergence Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198 7 Identification of Fractional Order Linear and Nonlinear Systems from Experimental or Simulated Data. . . . . . . . . . . . . . . . . 199 7.1 Fractional Order System (FOS) Identification using TFs . . . . . 201 7.1.1 Linear FOS Identification. . . . . . . . . . . . . . . . . . . . . . . . . 201 7.1.2 Nonlinear FOS Identification . . . . . . . . . . . . . . . . . . . . . 204
Contents
xi
7.2
Simulation Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206 Case study 7.1: Identification of Linear Single Input Single Output (SISO) FOS . . . . . . . . . . . . . . . . . . . . . . .206 Case study 7.2: Identification of Linear SISO Integer Order System (IOS). . . . . . . . . . . . . . . . . . . . . . . . . . . . .207 Case study 7.3: Identification of Linear Multi-Input Single Output IOS . . . . . . . . . . . . . . . . . . . . . . . .209 Case study 7.4: Identification of Nonlinear SISO FOS . . . . . . . .214 Case study 7.5: Verification of applicability of proposed identification method for sinusoidal signal, square wave signal, Sawtooth wave signal, step signal, pseudo random binary signal . . . .216 7.3 MATLAB Codes for Simulation Examples . . . . . . . . . . . . . . . . . 218 7.4 Summary of Chapter Deliverables. . . . . . . . . . . . . . . . . . . . . . . . 232 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 8 Design of Fractional Order Controllers using Triangular Strip Operational Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235 8.1 Triangular Strip Operational Matrices–Based Fractional Order Controller Design Method. . . . . . . . . . . . . . . . . . . . . . . . . 237 8.2 Constrained Nonlinear Optimization . . . . . . . . . . . . . . . . . . . . . 242 8.2.1 Luus-Jaakola (LJ) Multipass Optimization Method . . . 242 8.2.2 Particle Swarm Optimization Method . . . . . . . . . . . . . . 244 8.3 Simulation Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 8.3.1 Design of Robust Fractional PIλDµ Controller for a Heating Furnace System . . . . . . . . . . . . . . . . . . . . . 246 8.3.2 Design of Fractional Order PIλDµDµ2 Controller for Automatic Voltage Regulator System . . . . . . . . . . . 256 8.3.3 Design of Fractional Order PIλ Controller, Fractional PDµ Controller, Fractional Order PIλDµ Controller with Fractional Order Filter, and Series Form of Fractional Order PIλDµ Controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263 8.4 MATLAB Codes for Simulation Examples . . . . . . . . . . . . . . . . . 265 8.5 Summary of Chapter Deliverables. . . . . . . . . . . . . . . . . . . . . . . . 279 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280 9 Rational Integer Order System Approximation for Irrational Fractional Order Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283 9.1 The Proposed Integer-Order Approximation Method . . . . . . . 286 9.2 Simulation Example. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290 9.3 MATLAB Codes for Simulation Example. . . . . . . . . . . . . . . . . . 305 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
xii
Contents
10 Numerical Method for Solving Fractional-Order Optimal Control Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 10.1 The Proposed Numerical Method . . . . . . . . . . . . . . . . . . . . . . . . 312 10.2 Simulation Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315 Case study 10.1: Optimal control of linear time invariant integer order system (IOS) . . . . . . . . . . . . . . . .316 Case study 10.2: Optimal control of linear time-varying fractional-order system (FOS) . . . . . . . . . . . . .316 Case study 10.3: Optimal control of nonlinear FOS . . . . . . . . . .318 Case study 10.4: Optimal control of two-dimensional IOS . . . .320 10.3 MATLAB® Codes for Simulation Examples. . . . . . . . . . . . . . . . 321 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
Preface This book discusses significant applications of triangular functions in simulation, identification, and control of fractional-order processes. Processes exhibiting fractional-order dynamics are called fractional-order processes. Fractional Calculus (FC) is an active branch of mathematical analysis that deals with the theory of differentiation and integration of arbitrary order. It is also known as Generalized Integral and Differential Calculus, and Calculus of Arbitrary Order. The concept of the fractionalorder derivative was first discussed by Leibniz and L’Hospital almost three hundred years ago (i.e., at the end of the seventeenth century), a time when the foundations of the integer order calculus were developed by Isaac Newton and Gottfried Wilhelm Leibniz. Leibniz introduced the symbol dn f ðxÞ=dxn , n2N, to denote the nth derivative of a function f ðxÞ. In Leibnizs’ letter to Guillaume de l’Hospital, dated 30 September 1695 (which is considered as the date of birth of fractional calculus), he raised the question about the possibility of generalizing the operation of classical differentiation to noninteger orders. This question aroused l’Hospital’s inquisitiveness and he replied to Leibniz with another question: “What does dn f ðxÞ=dxn if n ¼ 1=2 mean?” Leibniz replied, “It will lead to a paradox from which one day useful consequence will be drawn.” L’Hosptial’s curiosity about the meaning of the semiderivative (i.e., dn f ðxÞ=dxn , n ¼ 1=2, a fraction or rational number) gave rise to the name of this subject (FC), and its name has remained the same, even though n can be any real number (rational or irrational). Although the same name is used throughout this book due to historical reasons, it should be understood to be noninteger (arbitrary real number) calculus, to be exact. As a matter of fact, even complex numbers can be allowed. Since the inception of FC, many great mathematicians (pure and applied), such as N. H. Abel, M. Caputo, L. Euler, A. K. Grunwald, J. Fourier, J. Hadamard, G. H. Hardy, O. Heaviside, H. J. Holmgren, P. S. Laplace, G. W. Leibniz, A. V. Letnikov, J. Liouville, B. Riemann, M. Riesz, and H. Weyl have contributed to this field. FC remained unknown to many applied mathematicians, scientists, and engineers because several definitions formulated for fractional-order derivative worked only in some situations. The mathematical theory of the subject seemed very different from that of integer order calculus. FC was considered to be an abstract area involving only mathematical manipulation of little or no use, and was thought to have no applications. Almost three decades ago, the mathematics and applied mathematics fraternity realized the potential of FC and started developing essential mathematical theory to establish it. Since then, FC has emerged as an important and efficient tool for the study of dynamical systems where xiii
xiv
Preface
classical calculus reveals strong limitations. The books and monographs of Oldham and Spanier (1974), Oustaloup (1991, 1994, 1995), Miller and Ross (1993), Samko, Kilbas, and Marichev (1993), Kiryakova (1994), Carpinteri and Mainardi (1997), Podlubny (1999), and Hilfer (2000) have been instrumental in introducing FC to the pure and applied mathematics community. FC has been applied in diffusion processes, modeling of the mechanical properties of materials, signal processing, advection and dispersion of solutes in natural porous or fractured media, image processing, modeling of the behavior of viscoelastic and viscoplastic materials under external influences, pharmacokinetics, bioengineering, description of mechanical systems subject to damping, relaxation, and reaction kinetics of polymers, ultraslow processes, connections to the theory of random walks, finance, control theory, and psychology. FC has substantial applications within the various fields of mathematics itself. One of the major advantages of FC is that it includes the integer order calculus as a special case (i.e., a superset of integer order calculus). Therefore, FC can accomplish what its counterpart cannot achieve, especially capturing the memory and heredity of a process. FC is a useful and efficient tool to reveal many phenomena in nature because nature has memory. We believe that FC will be the only type of calculus in the future. There are too many books available on the subject of FC. However, these books separate out topics in a way that can be confusing to students and less experienced researchers. There are some books addressing the “pure mathematical” side of the problems without taking into consideration those questions that arise in the applications mentioned earlier, and some that present the engineer’s point of view without rigorous mathematical justification. Whereas a unified approach to address a variety of topics starting from theory to application along with source codes being available is very much desirable for less experienced researchers, it will save their time to be invested toward further advancements. All of our struggles at one point of time as beginners and our efforts to make a way out gave us an invaluable experience, one that impelled us to write this book. An important goal of this book is to employ triangular orthogonal functions and triangular strip operational matrices to devise new numerical methods for simulation, identification, and control of fractional-order processes. The use of orthogonal functions as bases of expansion for squared integrable real-valued functions is a standard method in mathematical analysis and computational techniques. Numerous sets of orthogonal bases functions are available in mathematics. The existing sets of orthogonal functions can be categorized into two classes. The first class includes the classical sets of continuous functions such as sine-cosine functions, Legendre, Laguerre, Chebyshev, Jacobi, and Hermite orthogonal functions, and so on, which are continuous over their intervals of definition and consequently are well suited to approximate continuous functions. The second class consists of piecewise constant functions with
Preface
xv
inherent discontinuities. Walsh, block pulse, and Haar functions fall under the second category. The triangular orthogonal function sets, which are the foundation of most of numerical methods formulated in this book, are a complementary pair of piecewise linear polynomial function sets evolved from a simple dissection of block pulse function (BPF) set. The reason for choosing orthogonal functions as basis of the numerical methods presented in this book is that they can reduce the calculus of continuous dynamical systems to an attractive algebra, that is, they can convert integral, integrodifferential, differential, differential-algebraic, and partial differential equations into a set of algebraic equations. The triangular functions are only selected among the existing enormous orthogonal functions because it is much easier to work with them than with others. This book is organized into 10 chapters. The objectives, original contributions, and key findings of each chapter are summarized in the following paragraphs. Chapter 1 is the backbone of the book, as it contains all of the mathematical postulations used in the subsequent chapters of the book. The special mathematical functions that play a vital role in establishing the proper definition of operators of fractional calculus are briefly reviewed. The most widely used definitions and properties of fractional-order integrals and fractional-order derivatives are provided. To analyze the behavior of linear lumped fractionalorder systems, the Laplace transforms of fractional-order operators are derived, and then fractional-order systems are categorized along with a discussion about their stability. Different types of fractional-order controllers as well as advantages and disadvantages of the classical types are discussed. Triangular orthogonal functions are presented along with approximation formulae for estimating functions and fractional-order integrals. Finally, triangular strip operational matrices, which are the basis of the proposed robust controller tuning technique in Chapter 8, are derived from the classical finite difference formula and the Grunwald-Letnikov fractional-order derivative. Source codes developed in MATLAB® for triangular function operational matrices and triangular strip operational matrices are provided. Weakly singular integral equations are extremely difficult to solve. In Chapter 2, an effective numerical method is developed using triangular orthogonal functions to solve weakly singular (WS) Fredholm, WS Volterra, and WS Volterra-Fredholm integral equations. Mathematical theory regarding the existence of unique solutions to weakly singular Volterra-Fredholm integral equations is provided. The convergence of the approximate solution to the actual solution is studied theoretically and numerically. The proposed method is tested on a set of test problems and then applied to physical process models described by weakly singular integral equations. Abel’s integral equation, one of the very first integral equations, was seriously investigated by Niels Henrik Abel in 1823 and by Liouville in 1832 as a fractional power of the operator of antiderivation. This equation is encountered in the inversion of seismic travel times, stereology of spherical
xvi
Preface
particles, spectroscopy of gas discharges (more generally, “tomography” of cylindrically or spherically symmetric objects like, e.g., globular clusters of stars), and determination of the refractive index of optical fibers and electrochemistry. In Chapter 3, a novel numerical method using triangular orthogonal functions is developed to solve Abel’s integral equation (fractional-order integral equation) of the first and second kind. It is proved that Abel’s integral equations considered in this chapter have a unique solution in the given interval. The convergence analysis is carried out theoretically and numerically to prove that the proposed numerical algorithm can offer accurate approximate solutions that are very close to the true solutions of Abel’s integral equations under consideration, provided that a relatively small step size is employed. A wide variety of Abel’s integral equations is solved to demonstrate the applicability, accuracy, and stability of the proposed numerical algorithm. Encouraged by this success on the test problems, the proposed numerical method is applied to solve problems in electrochemistry, which are modeled by Abel’s integral equations. The obtained results confirm the practical appropriateness of the numerical algorithm for applications of Abel’s integral equations. Integro-differential equations of fractional order find their applications in heat transfer, thermodynamics, electrical conduction of polymers, and many more. It is well known that most of physical process models involving fractional-order integro-differential equations do not have exact solutions. This fact has been the driving force for numerous researchers to conduct research toward the development of efficient numerical methods to simulate such physical process models. The objective of Chapter 4 is to propose a novel numerical method based on triangular orthogonal functions for the numerical solutions of fractional-order integro-differential equations such as Fredholm integro-differential equations of fractional order, Volterra integro-differential equations of fractional order, and Fredholm-Volterra integro-differential equations of fractional order. It is theoretically shown that there exists a unique solution to the general form of the system of fractional-order integro-differential equations considered in this chapter. Convergence analysis is conducted to prove that in the limit of step size tends to zero, the proposed numerical method ensures the convergence of the approximate solution to the exact solution of fractionalorder integro-differential equations considered. Numerical examples as well as physical process models involving fractional-order integro-differential equations are solved to demonstrate the effectiveness of the proposed numerical method. The development of a single numerical method that is able to solve different forms of fractional-order differential equations and fractionalorder differential-algebraic equations is the prime objective of Chapter 5. Before construction of the numerical method, it is shown that the general form of system of fractional-order differential equations encompassing the aforesaid different forms has a unique solution in the given interval.
Preface
xvii
Convergence analysis is carried out to show that the approximate solution obtained by the proposed method can approach the original solution as the step size decreases to zero. The proposed method is applied to physical process models such as the Bagley-Torvik equation, the two-point BagleyTorvik equation, the plant-herbivore model, the computer virus model, the chemical Akzo Nobel problem, Robertson’s system describing the kinetics of autocatalytic reaction, and the high irradiance response of photo morphogenesis. In addition to the proposed method, the most popular semianalytical techniques such as the Adomian decomposition method (ADM), the homotopy analysis method (HAM), and the fractional differential transform method with Adomian polynomials (FDTM) are implemented as well on physical process models involving a stiff system of differential equations or stiff differential-algebraic equations. It is astonishing to note that ADM, HAM, and FDTM fail to simulate those process models even in the neighborhood of the initial time point 0, although they have successfully simulated many other physical process models. By contrast, the proposed method is able to produce valid approximate solution not only in the vicinity of the initial time point 0 but also in the desired time interval, which can be quite a bit larger than [0, 1]. In Chapter 6, the triangular functions–based numerical method is formulated to simulate fractional diffusion-wave equation. It is theoretically proved that the proposed method converges the approximate solution to the original solution of fractional diffusion-wave equation in the limit of step size tends to zero. Because the geometric and physical interpretation of fractional calculus is not as distinct as integer calculus, it is difficult to model real systems as fractional-order systems directly based on mechanistic analysis. Therefore, system identification is a practical way to model a fractional-order system, using experimental or simulated process data. Chapter 7 introduces an arbitrary order (note that the order can be integer or noninteger) system identification method based on the triangular orthogonal functions to estimate parameters including arbitrary differential orders and initial conditions of a model under consideration from experimental or simulated data. Five identification problems encompassing integer- and nonintegerorder linear and nonlinear systems are given to validate the proposed method. It is proved that the proposed method works well for various kinds of input excitation signal such as step signal, pseudo random binary signal, square wave signal, Sawtooth wave signal, and pulse signal. The obtained results are compared with the results of some of the existing numerical methods, and it is found that the proposed method demonstrates superior performance over those methods. To tackle plant uncertainty issues, many controller design methods are developed. The convenient one among these methods is designing a robust fractional PI λ Dμ controller. Chapter 8 proposes a simple tuning technique aimed to produce a robust noninteger order PID controller exhibiting iso-
xviii
Preface
damping property during the reparameterization of a plant. The required robustness property is achieved by allowing the fractional PID control system to imitate the dynamics of a reference system with Bode’s ideal transfer function in its forward path. The objective of designing robust controller by tracking the dynamics of reference control system is defined mathematically as an H∞ -optimal control problem. Fractional differential systems are transformed into algebraic equations by the use of triangular strip operational matrices. The H∞ -optimal control problem is then changed to an ∞-norm minimization of a parameter (KC ; KI ; Kd ; λ ; μ) varying square matrix. Global optimization techniques; Luus-Jaakola direct searche, and particle swarm optimization are employed to find the optimum values of fractional PID controller parameters. The proposed method of control system design is implemented in heating furnace temperature control, automatic voltage regulator systems, and some integer and fractional-order process models. Fractional PIλ, fractional PDµ, PIλDµDµ2, fractional PID with fractional-order filter, and the series form of fractional PID controller are designed as optimal controllers using the triangular strip operational matrix–based control design method. The performance of the proposed fractional-order controller tuning technique is found to be better than the performance of some fractional-order controller tuning methodologies reported in the literature. Triangular strip operational matrices proposed from the perspective of mathematics (for the solution of fractional differential and partial differential equation) finds its elegant application in the proposed method of control system design. Chapter 9 presents a new numerical method to find an equivalent finite dimensional integer order system for an infinite dimensional fractionalorder system. The approximated rational integer order system owns characteristics close enough to that of irrational fractional-order system that can be used in place of the original fractional-order system. In comparison to Crone, Carlson, Charef, Matsuda, and continued fraction expansion approximation methods, the proposed method offers lower order rational approximation that precisely mimics the dynamics of the original irrational fractional-order system. One significant attribute of the proposed method, which none of the abovementioned methods possess, is that the order of rational approximation can be controlled while obtaining accurate approximation to the fractional-order system. Optimal control problems arise naturally in various areas of science, engineering, and mathematics. Considerable work has been done in the area of integer optimal control problems (IOCPs), whose dynamics are described by conventional integer differential equations. Recently, it has been demonstrated that fractional differential equations are more accurate than integer differential equations to describe the dynamic behavior of many real-world processes. Fractional optimal control problems (FOCPs) are a subclass of classical optimal control problems whose dynamics are described by fractional differential equations. It is well known that the analytical
Preface
xix
solutions of FOCPs generally do not exist except for special cases, and therefore, numerical methods to obtain an approximate solution have become the preferred approach for solving FOCPs. An effective numerical method is devised in Chapter 10 to solve FOCPs. The proposed method is implemented on a set of examples comprising linear time invariant optimal control problems, time-variant fractional-order optimal control problems, nonlinear fractional-order optimal control problems, and two-dimensional integer order optimal control problems. Comparisons between the obtained results and that of some of the existing numerical methods shows the superiority of the proposed method over most of the others. This book may be treated as an interdisciplinary reference book, and it also may be used as a textbook for courses related to numerical methods for fractional-order systems, fractional-order process identification, fractional-order controller design, and fractional-order optimal control. The book is suitable for engineering and basic science researchers and scientists irrespective of their disciplines. The broader scope of the book makes it suitable for the budding researcher. Final year mechanical, electrical, chemical, mathematics, physics, and biomedical graduate students will find support when consulting this book for research-oriented courses. The reader is expected to be acquainted with classical calculus (differential and integral calculus, elementary theory of integral, differential, integro-differential, and partial differential equations), process identification and control, and optimal control. After reading Chapter 1, the reader will be able to understand the concepts presented in the subsequent chapters. The extensive literature survey on each concept addressed in the book is deliberately avoided. Every chapter begins with the necessary basic information and adequate background for the reader to grasp the concepts discussed in the chapter. Source codes developed in MATLAB are provided in each chapter, allowing the interested reader to take advantage of these codes to broaden and enhance the scope of the book itself as well as develop new results. We hope that readers will find this book useful and valuable in the advancement of their knowledge and their field. We look forward to receiving comments and suggestions from researchers, pure and applied mathematicians, scientists, and engineers. India, April 2018
Seshu Kumar Damarla Madhusree Kundu
MATLAB® is a trademark of The MathWorks, Inc. and is used with permission. The MathWorks does not warrant the accuracy of the text or exercises in this book. This book’s use or discussion of MATLAB® software or related products does not constitute endorsement or sponsorship by The MathWorks of a particular pedagogical approach or particular use of the MATLAB® software.
This page intentionally left blank
Acknowledgments We are grateful to the reviewers: Dr. Sillas Hadjiloucas (Department of Bio-engineering, University of Reading), Mr. Amr Elsonbaty (Engineering Mathematics and Physics Department, Faculty of Engineering, Mansoura University, Egypt), Dr. YangQuan Chen (University of California Merced), and three anonymous reviewers for their valuable suggestions that helped to enhance the quality of the book. We would like to thank the CRC press team, in particular, Dr. Gagandeep Singh, Senior Editor (Acquisitions) for engineering/environmental sciences, and CRC Press, for their patience, encouragement, invaluable suggestions, and keeping trust in us during the process of writing this book. We are thankful to HOD, Chemical Engineering, NIT Rourkela, and Director, NIT Rourkela for their kind cooperation during preparation of the book. Our sincere thanks go to those who directly or indirectly helped us during preparation of the book. Last but not least, we thank our families for their unconditional love and support.
xxi
This page intentionally left blank
About the Authors Seshu Kumar Damarla was born in the year 1985 in Chirala, Prakasam, Andhra Pradesh, India. He did his B.Tech (Chemical Engineering) from Bapatla Engineering College, Bapatla, Andhra Pradesh, India (2008), and M.Tech (Chemical Engineering) from NIT Rourkela, Odisha, India (2011). Mr. Damarla submitted his Ph.D dissertation (Developing Numerical Methods for Simulation, Identification and Control of Fractional Order Process) to NIT Rourkela, Odisha, India (2017). Mr. Damarla served as an Assistant Professor for a short duration (from 5th August 2011 to 31st December 2011) in Department of Chemical Engineering at Maulana Azad National Institute of Technology Bhopal, Madhya Pradesh, India, and for a year (from July 2017 to July 2018) in Department of Chemical Engineering at C.V. Raman College of Engineering, Bhubaneswar, Odisha, India. Mr. Damarla has published a couple of research articles in the internationally refereed journals to his credit and also published in the proceedings of national and international conferences. Mr. Damarla co-authored a reference textbook Chemometric Monitoring: Product Quality Assessment, Process Fault Detection, and Applications (CRC Press). Mr. Damarla has been a referee for Acta Biotheoretica, Journal of King Saud Science, and Applied and Computational Mathematics. Mr. Damarla is a member of International Association of Engineers (IAENG), Fractional Calculus and Application Group, and Allahabad Mathematical Society. Madhusree Kundu started her academic pursuits with a graduation in chemistry, with honors (University of Calcutta) followed by graduation and post-graduation in chemical engineering from the Rajabazar Science College, University of Calcutta, (1990–1992). Dr. Kundu gained experience as a process engineer at Simon Carves (I) Ltd., Kolkata (1993–1998). In the next phase of her scholarly pursuit, Dr. Kundu earned her Ph.D from the Indian Institute of Technology, Kharagpur (1999–2004), and started her academic profession as the faculty of the Chemical Engineering Group, BITS Pilani, Rajasthan (2004–2006). She joined the NIT Rourkela in 2007 and is continuing there as Professor in the Department of Chemical Engineering. Apart from teaching, she has focused her research activities in chemometrics along with fractional order process modeling and control, solution thermodynamics, and fluid-phase equilibria. Dr. Kundu has authored several research articles in International refereed journals and has a few book chapters, and a reference text book (Chemometric Monitoring: Product Quality Assessment, Process Fault Detection, and Applications, CRC Press) to her credit.
xxiii
This page intentionally left blank
1 Mathematical Postulations
1.1 Special Functions In this section, the important functions that are essential for the formulation of fractional-order integral and fractional-order derivative, and that will be used in the following chapters, are provided. 1.1.1 Gamma Function [1] The gamma function is the generalization of the factorial function to noninteger numbers. It plays a vital role in defining fractional-order integrals and fractional-order derivatives. The complete gamma function is defined by the Euler limit as ΓðxÞ ¼
N!Nx lim N ! ∞ xðx þ 1Þðx þ 2Þ . . . . . . ðx þ NÞ
ð1:1Þ
and by the integral transform as ð∞
ΓðxÞ ¼ ey yx1 dy; x > 0:
ð1:2Þ
0
The definition in Equation (1.2) is more practicable than that in Equation (1.1), even though it is restricted to positive values of x. Following are some useful properties of the gamma function. The recurrence relationship given here is the most important property of the gamma function, and it can be obtained by applying the integration by parts to Equation (1.2):
1
2
Fractional Order Processes
Γðx þ 1Þ ¼ xΓðxÞ:
ð1:3Þ
For x ≤ 0, the gamma function in Equation (1.2) becomes infinite. But the ratios of the gamma functions of negative integers are, however, finite, as shown in the next equation; ΓðPÞ Q! ¼ ð1ÞQP ; where P; Q are arbitrary positive integers: ΓðQÞ P!
ð1:4Þ
The reflection of the gamma function ΓðxÞ is ΠcosecðΠxÞ Γ ð x þ 1Þ
ð1:5Þ
4x ΓðxÞΓ x þ 12 pffiffiffiffi : 2 Π
ð1:6Þ
ΓðxÞ ¼ and the duplication is Γð2xÞ ¼
The incomplete gamma function is ðc ∞ X cx xj ; yx1 expðyÞdy ¼ expðxÞ γ ðc; xÞ ¼ ΓðxÞ Γ ð j þ c þ 1Þ j¼0
ð1:7Þ
0
where γ ðc; xÞ is a finite single-valued analytic function of x and c. The recursive relation of γ ðc; xÞ is γ ðc 1; xÞ ¼ xγ ðc; xÞ þ
expðxÞ : ΓðcÞ
ð1:8Þ
1.1.2 Beta Function [1] The complete beta function, which is also known as Euler’s integral of the second kind, is defined by the beta integral as ð1 Bðp; qÞ ¼ yp1 ð1 yÞq1 dy; p > 0; q > 0: 0
ð1:9Þ
3
Mathematical Postulations
The following relationship between the complete beta function and the complete gamma function can be used if either p or q is nonpositive; if this is the case, the integral in Equation (1.9) diverges: Bðp; qÞ ¼
ΓðpÞΓðqÞ ; 8p; q: Γðp þ qÞ
ð1:10Þ
The incomplete beta function of argument x is defined by ðx Bx ðp; qÞ ¼ yp1 ð1 yÞq1 dy:
ð1:11Þ
0
1.1.3 Mittag-Leffler Function The following one-parameter Mittag-Leffler function, introduced by MittagLeffler in [2], is an essential function used in modelling physical processes with the help of the fractional calculus concepts: E α ðzÞ ¼
∞ X
zk : Γ ð α k þ 1Þ k¼0
ð1:12Þ
The classical exponential function can be acquired from Equation (1.12) if α ¼ 1. The two-parameter Mittag-Leffler function, which is equally important as Equation (1.12) in fractional calculus, is given in the next equation [3]: E α ;β ðzÞ ¼
∞ X
zk ; α ; β > 0: Γð α k þ β Þ k¼0
ð1:13Þ
1.1.4 Hypergeometric Function The generalized hypergeometric function, which embraces several analytical functions as specific or limiting cases, is defined as ∞ Γðb1 Þ……Γ bq X Γða1 þ kÞ……Γ ap þ k zk ; p F q a1 ; ……; ap ; b1 ; ……; bq ; z ¼ Γða1 Þ……Γ ap k¼0 Γðb1 þ kÞ……Γ bq þ k k!
ð1:14Þ where bi ’s are not nonpositive integers.
4
Fractional Order Processes
The series in Equation (1.14) converges for all z if p ≤ q, and for jzj < 1 if p ¼ q þ 1. It diverges for all nonzero z if p > q þ 1. For the special case of p ¼ 2 and q ¼ 1, the generalized hypergeometric function in Equation (1.14) produces the well-known Gauss hypergeometric function 2 F 1 ða; b; c; zÞ ¼
∞ ΓðcÞ X Γða þ kÞΓðb þ kÞ zk ; k! ΓðaÞΓðbÞ k¼0 Γðc þ kÞ
ð1:15Þ
which frequently arises in physical problems. The relation among the hypergeometric function, the incomplete gamma function, and the incomplete beta function is γ ðv; zÞ ¼
1 1 ez F ð1; v þ 1; zÞ; 1 F 1 ðv; v þ 1; zÞ ¼ Γðv þ 1Þ Γðv þ 1Þ 1 1
ð1:16Þ
Bτ ðx; yÞ ¼ x1 τx 2 F 1 ðx; 1 y; x þ 1; τ Þ ¼ x1 τx ð1 τÞy 2 F 1 ðx þ y; 1; x þ 1; τ Þ: ð1:17Þ
1.1.5 Error Function and Complementary Error Function The error function, which comes across in integrating normal distribution and is an entire function, is defined as ðz 2 2 erf ðzÞ ¼ pffiffiffiffi et dt: Π
ð1:18Þ
0
When the argument z becomes infinite, then erf ð∞Þ ¼ 1. The series form of the error function is ∞ 2 X ð1Þn z2nþ1 2 z3 z5 z7 z9 ¼ pffiffiffiffi z þ þ þ . . . . . . : ð1:19Þ erf ðzÞ ¼ pffiffiffiffi 3 10 42 216 Π n¼0 ð2n þ 1Þn! Π The error function can be expressed by means of the hypergeometric function: 3 1 3 1 2 1 erf ðxÞ ¼ 2Π =2 xex 1 F 1 1; ; x2 ¼ 2Π =2 x1 F 1 ; ; x2 : 2 2 2
ð1:20Þ
5
Mathematical Postulations
The complementary error function is an entire function defined by ðz ð∞ 2 2 2 t2 erfcðzÞ ¼ 1 erf ðzÞ ¼ 1 pffiffiffiffi e dt ¼ pffiffiffiffi et dt Π Π 0
ð1:21Þ
z
and its series asymptotic expansion is ! ! 2 2 ∞ ∞ X X ez ez n 1 35 …ð2n 1Þ n ð2nÞ! ð1Þ ð1Þ ¼ pffiffiffiffi 1 þ : erfcðzÞ ¼ pffiffiffiffi 1þ n ð2z2 Þ z Π z Π n!ð2zÞ2n n¼1 n¼1 ð1:22Þ 1.1.6 Bessel Functions Of all higher transcendental functions, the Bessel functions are possibly most omnipresent. They appear often in physical phenomena such as electromagnetic waves in a cylindrical waveguide, pressure amplitudes of inviscid rotational flows, heat conduction in a cylindrical object, diffusion problems on a lattice, dynamics of floating bodies, signal processing, and so on. The Bessel functions of the first kind are the solutions to the Bessel differential equations x2
d2 y dy þ x þ x2 n2 y ¼ 0; 2 dx dx
ð1:23Þ
which are finite at the origin x ¼ 0. The Bessel functions of the first kind, which are also called cylindrical functions or cylindrical harmonics, are defined by infinite series expansion as J n ð xÞ ¼
x 2mþn ð1Þm ; n is the order of the Bessel function: m!Γðm þ n þ 1Þ 2 m¼0 ∞ X
ð1:24Þ The Bessel functions of the second kind defined here, occasionally called Weber functions or Neumann functions, are the solutions to Equation (1.23) when it becomes singular at x ¼ 0: Yn ðxÞ ¼
Jn ðxÞ cosðnΠÞ Jn ðxÞ : sinðnΠÞ
ð1:25Þ
6
Fractional Order Processes
1.2 Definitions and Properties of Fractional-Order Operators In this section, the notion of generalizing classical calculus to derive definitions of fractional-order integral and fractional-order derivative is presented [4]. 1.2.1 Grunwald-Letnikov (GL) Fractional-Order Derivative Extrapolating the applicability of classical backward difference formula to derivative of non-integer order gives rise to the formation of the Grunwald-Letnikov (GL) fractional-order derivative. Perhaps, GL is the first definition proposed for differentiation of noninteger order. Let us consider a real function f ðtÞ (t2½0; b). The function is said to be in the space Cμ ; μ2R if there exists a real number pð> μÞ, such that f ðtÞ ¼ tp f1 ðtÞ; where f1 ðtÞ2C½a; ∞Þ; and it is said to be in the space Cnμ if and only if f ðnÞ 2Cμ , n2Ν. We now express the nth order derivative (n is an integer) of the casual function f ðtÞ (i.e., f ðtÞ ¼ 0, t < 0) in terms of backward difference formula: N X dn f ðtÞ rn f ðtÞ i n n f ðt ihÞ ; ffi ¼h ð1Þ i dtn hn i¼0
ð1:26Þ
n! n where ¼ , h ¼ ðb aÞ=N, N is the total number of equidistant i i!ðn iÞ! nodes in the interval ½0; b. Equation (1.26) is the form of nth order derivative of the discretized n function, f ðtÞ. Because becomes zero for all values of i greater than n, i the upper limit of summation in this definition can be increased to infinity. Rewrite Equation (1.26) as GL n n 0 D t f ðtÞ ffi h
∞ X i¼0
d n ð1Þ f ðt ihÞ ; D ¼ : i dt i
ð1:27Þ
The following definition for the Grunwald-Letnikov fractional-order derivative can be obtained by putting α in place of n in Equation (1.27): GL α α 0 D t f ðtÞ ffi h
∞ X i¼0
ð1Þi
α f ðt ihÞ : i
ð1:28Þ
7
Mathematical Postulations
When the sign of α is negative, Equation (1.28) turns out to be a fractionalorder integral. The integral transform definition of GL fractional-order derivative is [5] ð n1 X f ðkÞ ð0Þt α þk 1 þ ¼ ðt τÞn α 1 f ðnÞ ðτÞdτ; Γ ð α þ k þ 1 Þ Γ ð n α Þ k¼0 t
GL α 0 D t f ðt Þ
ð1:29Þ
0
where n 1 ≤ α < n, n2Zþ , t > 0. The reason why the derivative possesses nonlocal prop fractional-order α erty is that the term in Equation (1.28) will never become zero; that is, i determining the fractional-order derivative of any function requires its entire history. Therefore, it needs infinite memory and thus is more suitable to explain long memory processes mathematically. It is worth mentioning here that classical calculus is a particular case of the fractional calculus. The GL fractional-order derivative in Equation (1.29) is the left fractional-order derivative, because the lower terminal of the fractional integral is fixed at the left end of the interval ½0; b and the upper terminal moves in the interval. If the upper terminal of the fractional integral in Equation (1.29) is fixed at the right end of the interval ½0; b and the lower terminal is moving, then the GL fractional derivative is called the right fractional derivative. Let us suppose that the independent variable t is time and the function f ðtÞ describes the dynamic behavior of a process. If τ < t (t is the current instant), then the past of this process can be described by the state f ðτÞ. If τ > t, then the state f ðτ Þ belongs to the future of the process. Therefore, the left fractional derivative is only considered throughout this book as it requires the past information of the process to describe it mathematically. 1.2.2 Riemann-Liouville (RL) Fractional-Order Integral Cauchy’s formula for repeated integration, which reduces n-fold integration of function f ðtÞ to single integral, is:
f
n
ðtÞ ¼
n 0J t
ðt 1 f ðt Þ ¼ ðt xÞn1 f ðxÞdx; n1 0
where n is a positive integer. Equation (1.30) can be written as
ð1:30Þ
8
Fractional Order Processes
n 0J t
ðt 1 f ðt Þ ¼ ðt xÞn1 f ðxÞdx; ΓðnÞ
ð1:31Þ
0
where ΓðnÞ is a well-known Euler’s Gamma function: ΓðnÞ ¼
Ð∞
ex xn1 dx.
0
Equation (1.31) permits us to replace n with α to obtain a fractional-order integral: α 0J t
ðt 1 f ðt Þ ¼ ðt xÞ α 1 f ðxÞdx: Γð α Þ
ð1:32Þ
0
1.2.3 Riemann-Liouville Fractional-Order Derivative The left Riemann-Liouville (RL) fractional-order derivative of function f ðtÞ is defined as RL α 0 D t f ðtÞ
n n α
¼D J
ðt 1 dn f ðtÞ ¼ ðt τÞn α 1 f ðτÞdτ; t > 0; Γðm α Þ dtn
ð1:33Þ
0
where α is a noninteger that satisfies the relation n 1 < α ≤ n; n2Zþ . The right Riemann-Liouville fractional-order derivative is RL α 0 D t f ðtÞ
n n α
¼D J
ðb 1 dn f ðtÞ ¼ ðt τÞn α 1 f ðτÞdτ; t < b: Γðn α Þ dtn
ð1:34Þ
t
1.2.4 Caputo Fractional-Order Derivative As we shall show in the next subsection, Riemann-Liouville fractional differential equations lack widespread physical applications because of the need for fractional-order initial conditions. To enable fractional calculus concepts to be applied in different applied branches of science and technology, Caputo modified Equation (1.34) as shown in the following definition [6]. The left Caputo fractional-order derivative is
C α 0 D t f ðtÞ
¼J
ðt 1 f ðtÞ ¼ ðt τÞn α 1 f n ðτÞdτ; t > 0 Γðn α Þ
n α n
0
ð1:35Þ
9
Mathematical Postulations
and the right Caputo fractional derivative is
C α 0 D t f ðtÞ
¼J
ðb 1 f ðtÞ ¼ ðt τÞn α 1 f n ðτÞdτ; t < b: Γðn α Þ
n α n
ð1:36Þ
t
When there are homogenous initial conditions, Riemann-Liouville and Caputo fractional-order derivatives are equivalent. The definitions of operators of fractional calculus presented so far will become the operators of classical calculus when the fractional-order is equal to an integer. Like operators of classical calculus, fractional-order operators also have a physical interpretation. Fractional-order integrals can be understood as the area under a shape changing curve, whereas fractional-order derivatives imply the integer order derivative of area under the shape-changing curve [7]. 1.2.5 Properties of GL, RL, and Caputo Fractional-Order Derivatives Some useful properties of fractional-order operators that we shall use in the following chapters are provided here [4, 5]. For f ðtÞ2Cμ , μ > 1 and n 1 ≤ α < n, p 1 ≤ β < p, p; n; q2Zþ , α ; β2Rþ : • • • •
α β 0 J t 0 J t f ðtÞ
¼ 0 J βt 0 J αt f ðtÞ ¼ 0 J αþβ f ðt Þ t property).
(semigroup
GL α 0 Dt c
¼ ðct α Þ=Γð1 α Þ, RL0 D αt c ¼ ðct α Þ=Γð1 α Þ, C0 D αt c ¼ 0, c is a constant. ! n1 k X t ðkÞ C α RL α f ð 0Þ . f ðtÞ 0 D t f ðt Þ ¼ 0 D t k! k¼0
• •
f ðsÞ ð0Þ ¼ 0, s ¼ n; n þ 1; …; q.
q q RL α RL α RL αþq f ðtÞ, 0 D t f ðt Þ ¼ 0 D t 0 D t D0;t f ðtÞ ¼ 0 D t
•
commutative
lim α !n 0 J αt f ðtÞ ¼ 0 J nt f ðtÞ (consistency property with the integer order integral).
¼ RL0 D αt 0 J αt f ðtÞ ¼ f ðtÞ holds for n ¼ 1. n1 k X t ðkÞ αC α f ð0Þ. J D f ð t Þ ¼ f ð t Þ 0 t0 t k! k¼0
q q C α αþq C α D D f ð t Þ ¼ D D f ð t Þ ¼ C0 D t f ðtÞ, 0 t 0 t 0 t 0 t
•
and
C α α 0 D t 0 J t f ðtÞ
f ðsÞ ð0Þ ¼ 0, s ¼ 0; 1; …; q.
10
Fractional Order Processes
8 < 0 Dαt 0 Dβt f ðtÞ ¼ 0 Dβt 0 Dαt f ðtÞ ≠ 0 Dtαþβ f ðtÞ
, : 0 Dα 0 Dβ f ðtÞ ≠ 0 Dβ 0 Dα f ðtÞ ¼ 0 Dαþβ f ðtÞ t t t t t
γ RL γ C γ 0 D t 2½ 0 D t ; 0 D t ; γ2½ α ; β.
1.3 Laplace Transforms of Fractional-Order Operators Before obtaining Laplace transforms of fractional-order integrals and fractional-order derivatives, let us briefly review the basics of the Laplace transform. Let us suppose the function, f ðtÞ. The Laplace transform of f ðtÞ is defined by ð∞
est f ðtÞdt: L f t ¼ FðsÞ ¼
ð1:37Þ
0
This definition is valid if and only if the function f ðtÞ is of exponential order β, that is eβt jf ðtÞj ≤ M, 8t > δ , where M, δ are positive constants. The original function f ðtÞ can be retrieved from FðsÞ by using the following inverse Laplace transform: ð cþia
1 lim est FðsÞds; c ¼ ReðsÞ > c0 ; f ðtÞ ¼ L1 FðsÞ ¼ 2Πi a ! ∞ cia
ð1:38Þ
where c0 lies in the right half plane of the absolute convergence of Equation (1.37). Let us assume two functions f1 ðtÞ and f2 ðtÞ ( f 1 ðtÞ ¼ f2 ðtÞ ¼ 0; t < 0). The convolution of f1 ðtÞ and f2 ðtÞ creates a function f3 ðtÞ according to the following equation: f3 ðtÞ ¼ f1 ðtÞ f2 ðtÞ ¼
ðt 0
f1 ðt τÞf2 ðτÞdτ ¼
ðt
f1 ðτ Þf2 ðt τÞdτ:
ð1:39Þ
0
The mathematical operation in Equation (1.39) is valuable in mathematical physics and probability theory.
11
Mathematical Postulations
The Laplace transform of Equation (1.39) is L f 3 ðtÞ ¼ L f 1 ðtÞ f2 ðtÞ ¼ L f 1 ðtÞ L f 2 ðtÞ ¼ F1 ðsÞF2 ðsÞ:
ð1:40Þ
Equation (1.39) will be used in the derivation of the Laplace transform of the Riemann-Liouville fractional-order integral. For the Laplace transform of the fractional-order derivative, we need the following equation: n1 n1 X X snk1 f ðkÞ ð0Þ ¼ sn FðsÞ sk f ðnk1Þ ð0Þ: L Dn f ðtÞ ¼ sn FðsÞ k¼0
ð1:41Þ
k¼0
Let us recall the Riemann-Liouville fractional-order integral: α 0 J t f ðt Þ ¼
1 Γð α Þ
ðt
ðt τÞ α 1 f ðτ Þdτ ¼
0
t α 1 f ðtÞ : Γð α Þ
ð1:42Þ
Following Equation (1.39), L
α 0 J t f ðt Þ
¼
FðsÞ 1 1 L t α 1 f ðtÞ ¼ L t α 1 L f ðtÞ ¼ α : Γð α Þ Γð α Þ s
ð1:43Þ
By using Equations (1.40) and (1.42), we can get the Laplace transform of the Riemann-Liouville fractional-order derivative: n1 X sk 0 Dtnk1 0 Jtnα f ðtÞ t¼0 ; L RL0 Dαt f ðtÞ ¼ L 0 Dnt 0 Jtnα f ðtÞ ¼ sn L 0 Jtnα f ðtÞ k¼0
n1 h i FðsÞ X ðnk1Þ nα sk 0 Jt f ðtÞ ; ¼ sn nα 0 Jt t¼0 s k¼0
¼ sα FðsÞ
n1 h i X ðαk1Þ sk 0 Jt f ðtÞ k¼0
¼ sα FðsÞ
t¼0
;
n1 X sk RL0 Dtαk1 ð f ðtÞÞ t¼0 ; n 1 α5n; n 2 Zþ : k¼0
ð1:44Þ Similarly, the Laplace transform of the Caputo fractional-order derivative is
12
Fractional Order Processes
C α nα n L 0 Dnt f ðtÞ L 0 Dt f ðtÞ ¼ L 0 Jt 0 Dt f ðtÞ ¼ ; snα n1 h i X snk1 f ðkÞ ðtÞ ¼ snα sn FðsÞ k¼0
¼ sα FðsÞ
n1 X
h i sαk1 f ðkÞ ðtÞ
k¼0
t¼0
! t¼0
;
ð1:45Þ
; n 1 α5n; n 2 Zþ :
Equation (1.44) reveals that, in most scenarios, the RL fractional-order derivative does not work in physical process models, as it requires evaluation of the fractional-order derivative of f ðtÞ at the lower terminal t ¼ 0. There are very few situations in which these fractional initial conditions have a clear physical meaning [8]. From an application point of view, the Caputo form is more appropriate as it takes the values of function and its ðm 1Þ integer order derivatives at the lower terminal t ¼ 0. The Laplace transform of the GL fractional-order derivative can be obtained as shown here by using Equations (1.29) and (1.44): n1 h i X f ðkÞ ðtÞ t¼0 ðα þ kÞ! α αk1 ð kÞ þ s F ð s Þ s f ð t Þ ; t¼0 ðα þ k þ 1Þsαþkþ1 k¼0 k¼0
n1 X L GL0 Dαt f ðtÞ ¼
¼ sα FðsÞ; α 2 ½0; 1: ð1:46Þ
1.4 Fractional-Order Systems [9] Developing first principle models utilizing fractional-order integrals and/ or fractional-order derivatives for a linear lumped parameter SISO system results in the following fractional-order system: β
an GL0 Dαt n yðtÞ þ an1 GL0 Dαt n1 yðtÞ þ þ a0 GL0 Dαt 0 yðtÞ ¼ bn GL0 Dt n uðtÞ þ A;
P
n m P β bk GL0 Dt k uðtÞ ; ak GL0 Dαt k yðtÞ ¼
k¼0 β
ð1:47Þ
k¼0
β
where A ¼ bn1 GL0 Dt n1 uðtÞ þ þ b0 GL0 Dt 0 uðtÞ, ak , and bk are constants, α k and βk are real numbers, and yðtÞ and uðtÞ are the output and the input, respectively, of the system.
13
Mathematical Postulations
Assuming homogeneous initial conditions, a Laplace transform operation performed on Equation (1.47) gives the SISO linear time invariant fractional-order transfer function: GðsÞ ¼
YðsÞ bm s βm þ bm1 s βm1 þ . . . . . . þ b0 s β0 ¼ : UðsÞ an s α n þ an1 s α n1 þ . . . . . . þ a0 s α 0
ð1:48Þ
If the fractional orders on both sides of Equation (1.47) are integer multiples of the base order α , that is, α k ¼ βk ¼ k α , then the fractional-order system is said to be a commensurate-order system: m X
YðsÞ k¼0 ¼ n GðsÞ ¼ UðsÞ X
bk ðs α Þk ;
ð1:49Þ
α k
ak ðs Þ
k¼0
if not, then it is a noncommensurate order system. The commensurate-order system in Equation (1.49) can be considered as a pseudo-rational function, HðλÞ, of the variable λ ¼ s α : m X
HðλÞ ¼
bk λk
k¼0
n X
:
ð1:50Þ
ak λ
k
k¼0
If the order of commensurate-order systems is 1 q (q2Ζ þ ), then such a system is called a rational commensurate-order system. Similar to integer systems, the stability criteria for commensurate-order fractional systems can be defined as jargðλi Þj >
αΠ Π ; jargðλi Þj > ; 2 2q
ð1:51Þ
where λi ’s are poles of the commensurate-order system HðλÞ. For commensurate as well as noncommensurate order systems, the condition for bounded input-bounded output (BIBO) stability is lim jGðsÞj < M; M is a finite value: s!∞
ð1:52Þ
14
Fractional Order Processes
1.5 Fractional-Order PIλ, PDµ, and PIλDµ Controller [9] The output of the traditional PID controller is uðtÞ ¼ KC þ KI 0 J 1t eðtÞ þ Kd 0 D 1t eðtÞ:
ð1:53Þ
The transfer function form of the PID controller can be obtained by assuming zero initial conditions GC ðsÞ ¼ KC þ
KI þ Kd s: s
ð1:54Þ
The classical control actions—proportional, integral, and derivative—have positive as well as negative effects over the controlled system behavior. The proportional action increases the speed of the response and decreases the steady state error and the relative stability. The integral action eliminates the steady state error but decreases the relative stability. The derivative action increases the relative stability but makes the controlled system sensitive to high-frequency noisy signals. If the more general control actions of the form sp ; s1q ; p; q2Rþ are considered, the more acceptable tradeoffs between positive and negative effects can be accomplished. It will be noticed that the classical PID controller can meet only three performance specifications; however, if the order of the integrator and the differentiator can be an arbitrary order (integer and noninteger) and can be variable, the respective controller will bear five unknowns, that is, the controller gains (KC ; KI ; Kd ) and the order of the integrator and differentiator (p; q), and can achieve five control objectives, thus leading to the formulation of the generalized (fractional-order) PID controller. The ability of fulfilling more control objectives cannot be gained unless efficient tuning techniques are available. The following fractional-order integro-differential Equation expresses the output of the fractional-order PID controller: uðtÞ ¼ KC eðtÞ þ KI 0 J λt eðtÞ þ Kd GL0 D μt eðtÞ;
ð1:55Þ
where Jλ is the fractional integral of order λ, and Dµ is the GrunwaldLetnikov fractional derivative of order µ. Supposing null initial conditions and applying a Laplace Transform to the above equation, the transfer function of the fractional-order PID controller can be expressed by GC ðsÞ ¼
U ðs Þ KI ¼ KC þ λ þ Kd sμ ; λ; μ2R: s EðsÞ
ð1:56Þ
15
Mathematical Postulations
µ
µ=1
PD
P O
µ
PID
PI λ=1
PD
µ=1
λ
Classical PID
P O
PID
PI λ=1
λ
Fractional-order PID
FIGURE 1.1 Classical PID controller versus fractional-order PID controller
If λ ¼ 0 and μ ¼ 0, the fractional-order PID controller turns to be the traditional proportional controller. For λ ¼ 1 and μ ¼ 0, Equation (1.56) changes to the classical PI controller. The classical proportional derivative controller can be obtained from Equation (1.56) when λ ¼ 0 and μ ¼ 1. It is clear from Figure 1.1 that the classical P, PI, PD, and PID are the particular cases of the fractional-order PID controller. There are only four control configurations possible with PID controller, whereas numerous control structures can be obtained from the fractional-order PID controller by selecting the values of λ and μ in the Euclidean plane.
1.6 Triangular Orthogonal Functions The orthogonal triangular function (TF) sets developed by Deb et al. [10] are a complementary pair of piecewise linear polynomial function sets evolved from a simple dissection of block pulse function(BPF) set [10, 11]. The authors of [10, 11] derived a complementary pair of operational matrices for first-order integration in the TF domain and demonstrated that the TF domain technique for dynamical systems analysis is computationally more effective than the BPF domain technique. In this section, first, we review block pulse functions in brief and then we introduce the method of dissecting the block pulse function set to formulate a complementary pair of orthogonal triangular function sets. 1.6.1 Review of Block Pulse Functions Let us consider a square integral function f ðtÞ of Lebesgue measure, which is continuous in the interval ½0; T. Divide the interval into m subintervals of constant width h ¼ T=m as ½ti ; tiþ1 , i ¼ 0; 1; ……m 1.
16
Fractional Order Processes
Let ψm ðtÞ be a set of block pulse functions containing m component functions in the interval ½0; TÞ ψm ðtÞ ¼ ½ψ0 ðtÞ; ψ1 ðtÞ; ψ2 ðtÞ; ……; ψm1 ðtÞT1 m ;
ð1:57Þ
where ½. . .T signifies transpose. The ith component of the BPF vector ψm ðtÞ is defined as ψi ðtÞ ¼
1; 0;
ih ≤ t < ði þ 1Þh; ; i ¼ 0; 1; 2; ………; m 1: otherwise;
ð1:58Þ
The square integrable function f ðtÞ can be approximated by BPFs as f ðt Þ ¼
m1 X
fi ψðtÞ ¼ ½ f0 ; f1 ; ……; fm1 ψm ðtÞ ¼ FT ψm ðtÞ;
ð1:59Þ
i¼0
where the constant coefficients fi ’s are defined as fi ¼ 1h
ðiþ1 Ð Þh
f ðtÞdt.
ih
The BPF estimate for the first-order integration of f ðtÞ can be derived as [12] ðt Jf ðtÞ ¼ f ðτÞdτ ffi o
ðt
ðt F ψm ðτ Þdτ ¼ F T
ψm ðτÞdτ ¼ FT Pψm ðtÞ;
T
ð1:60Þ
o
0
where P is the operational matrix for 2 1 2 2 ... ... 60 1 2 2 ... 6 60 0 1 2 ... 6 .. domain: P ¼ h2 6 0 1 . 60 0 6. .. 4 .. . . . . . . ... . 0 0 0 ... 0
the first-order integration in the BPF 3 2 27 7 27 .. 7 . 7 .7 7 25 1 m m
The Riemann-Liouville fractional-order integral of f ðtÞ can be expressed by BPFs as [12] α T 0 J t f ðtÞ ffi F F α ψ m ðtÞ;
ð1:61Þ
17
Mathematical Postulations
where 2
1 60 6 60 6 . Fα ¼ 6 6 .. 6 6 40 0
ξ1 1 0 .. .
ξ2 ξ1 1 .. .
0 0
0 0
... ... ξ2 . . . ξ1 . . . .. .. . . .. 0 . 0 0
3 ξ m1 ξ m2 7 7 ξ m3 7 7 7; ξ k ¼ ðk þ 1Þ α þ1 2k α þ1 þ ðk 1Þ α þ1 ; k ¼ 1; 2; ……m 1: ξ m4 7 7 .. 7 . 5 1
1.6.2 Complementary Pair of Triangular Orthogonal Function Sets Let us divide the first component of the BPF vector ψm ðtÞ into a complementary pair of linear polynomial functions as shown in Figure 1.2. ψ0 ðtÞ ¼ T10 ðtÞ þ T20 ðtÞ;
ð1:62Þ
where T10 ðtÞ ¼ 1 ht and T20 ðtÞ ¼ ht , and the second component ψ1 ðtÞ, ψ1 ðtÞ ¼ T11 ðtÞ þ T21 ðtÞ;
ð1:63Þ
where T11 ðtÞ ¼ 1 th and T21 ðtÞ ¼ th h h . In the same fashion, we can divide the remaining components of ψm ðtÞ into respective complementary pairs of linear polynomial functions. Thus, for the whole set of BPFs, we now have two sets of linear polynomial functions, namely,T1m ðtÞ and T2m ðtÞ each contains m component functions in the interval ½0; T. ψm ðtÞ ¼ T1m ðtÞ þ T2m ðtÞ;
ð1:64Þ
where T1m ðtÞ ¼ ½T10 ðtÞ; T11 ðtÞ; ……; T1m1 ðtÞT and T2m ðtÞ ¼ ½T20 ðtÞ; T21 ðtÞ; ……; T2m1 ðtÞT .
The triangular function vectors; T1m ðtÞ and T2m ðtÞ together form the entire set of block pulse function set; hence, T1m ðtÞ and T2m ðtÞ complement each other as far as BPF set is considered. We recognize from Figure 1.2 that the shapes of T1i ’s and T2i ’s are left-handed and right-handed triangles, respectively. So, these two sets are named as the left-handed triangular function vector (LHTF) and the right-handed triangular function vector (RHTF), respectively. Now we define the ith component of the LHTF vector T1m ðtÞ as
18
Fractional Order Processes
ψ0(t)
T10(t)
1
T20(t)
+
1
0
h 0 ψ1(t)
0
h
T11(t)
h
0
2h
ψ2(t)
h
2h
h
2h
0
3h
+
2h
1
2h
3h
(m-1)h
t
mh
0
3h
T2m-1(t)
+
1
0
1
T1m-1(t)
ψm-1(t)
2h
T22(t) 1
0
1
+
T12(t)
1
h
T21(t) 1
1
0
1
t
(m-1)h mh
1
0
t
(m-1)h mh
FIGURE 1.2 Generation of TFs from BPFs
T1i ðtÞ ¼
8 < :
1
t ih ; ih ≤ t < ði þ 1Þh; ; i ¼ 0; 1; 2; ……; m 1 h 0; otherwise;
ð1:65Þ
and the ith component of the RHTF vector T2m ðtÞ as 8 < t ih ; ih ≤ t < ði þ 1Þh; T2i ðtÞ ¼ ; i ¼ 0; 1; 2; ……; m 1: h : 0; otherwise;
ð1:66Þ
19
Mathematical Postulations
Like BPFs, TFs can also be employed for the approximation of the square integrable function f ðtÞ in the interval ½0; T. ! m1
X lim f ðt Þ ¼ ci T1i ðtÞ þ di T2i ðtÞ : m!∞
ð1:67Þ
i¼0
Truncating the TFs series expansion in Equation (1.67) to m finite terms gives the practical approximation to the function, f ðtÞ2 L2 ð½0; TÞ, f ðtÞ≈
m1
X ci T1i ðtÞ þ di T2i ðtÞ ¼ CT T1m ðtÞ þ DT T2m ðtÞ;
ð1:68Þ
i¼0
where CT ¼ ½ c0 f ði þ1Þh .
c1
...
cm1 , DT ¼ ½ d0
d1
. . . dm1 , ci ¼ f ðihÞ, di ¼
The error analysis of estimating the function, f ðtÞ, in the TFs domain is thoroughly studied in [7]. The expressions for the coefficients; ci ’s and di ’s emphasis that the function evaluations at the equidistant nodes, ti , i ¼ 0; 1; 2; ……; m 1, are enough to find their numerical values. Whereas the coefficients in the BPF series representation in Equation (1.59) demand the integration of f ðtÞ. Thus, the function approximation in TF domain is computationally more effective compared to that in BPF domain. Because the members of T1m ðtÞ and T2m ðtÞ are mutually disjoint, their product can be expressed in the TF domain itself: 2
T10 ðtÞ 6 0 6 6 T T1ðtÞT1 ðtÞ ¼ 6 6 0 6 .. 4 . 0
0 T11 ðtÞ
0 0
0 .. .
T12 ðtÞ .. .
0
0
0 ... .. . .. . 0
0 0 0 0 T1ðm1Þ ðtÞ
3 7 7
7 7 ¼ diag T1ðtÞ ; 7 7 5 ð1:69Þ
2
T20 ðtÞ 6 0 6 6 T T2ðtÞT2 ðtÞ ¼ 6 6 0 6 .. 4 . 0
0 T21 ðtÞ
0 0
0 .. .
T22 ðtÞ .. .
0
0
0 ... .. . .. . 0
0 0 0 0 T2ðm1Þ ðtÞ
3 7 7
7 7 ¼ diag T2ðtÞ ; 7 7 5 ð1:70Þ
20
Fractional Order Processes
2
0 60 6 6 T1ðtÞT2T ðtÞ ¼ 6 0 6. 4 .. 0
0 0 0 .. . 0
0 0 0 ... . 0 .. .. .. . . 0 0
3 0 07 7 7 0 7 ¼ Om m ; 7 05 0
ð1:71Þ
where O is a null matrix. Each member of the LHTF vector, T1m ðtÞ, and the RHTF vector, T2m ðtÞ, possesses orthogonal property, hence, TFs are called orthogonal TFs. 8 h > < ; T1i ðtÞT1j ðtÞdt ¼ 3 > :h; 0 6
8 h T > < ; ifi ¼¼ j; ð ; T2i ðtÞT2j ðtÞdt ¼ 3 > :h; ifi≠j; 0 6
ðT
ifi ¼¼ j;
8i; j2 0; m 1
if i≠j; ð1:72Þ
Similar to the time function approximation by TFs as explained in Equation (1.68), any function, for example, G f ðtÞ , which can be linear or nonlinear, can be expanded into TFs:
G f ðtÞ ¼ ½ cn0
cn1
cn2
. . . cnm1 T1m ðtÞ þ ½ dn0
dn1
dn2
...
dnm1 T2m ðtÞ; ð1:73Þ
where ci ¼ G f ðti Þ , di ¼ ci1 . The product of square integrable functions h1 ðtÞ h2 ðtÞ . . . . . . hn ðtÞ can be approximated in the TF domain as h1 ðtÞ h2 ðtÞ . . . . . . hn ðtÞ ffi ½ c0
c1
… cm1 T1m ðtÞ þ ½ d0
d1
…
dm1 T2m ; ð1:74Þ
where ci ¼ h1 ð jhÞ h2 ð jhÞ . . . . . . hn ð jhÞ, di ¼ ciþ1 . 1.6.3 Expansion of Two Variable Function via TFs Let us consider the function in two variables, f ðx; tÞ2L2 ðJ1 J2 Þ, J1 ¼ ½0; T, J2 ¼ ½0; T. We now describe f ðx; tÞ by using the TFs series. The two intervals are split into an equal number (m) of subintervals using the constant step size, h, as described here: xi 2½ih; ði þ 1ÞhÞ; ti 2½ih; ði þ 1ÞhÞ; i2½0; m 1:
ð1:75Þ
21
Mathematical Postulations
By using Equation (1.68), 2
3 2 3 f ð0; tÞ f ðh; tÞ 6 7 f ðh; tÞ 6 7 f ð2h; tÞ 6 7 6 7 .. 6 7 6 7 . 6 7 . . 6 7 . 6 7 T 6 7: .. f ðx; tÞ ≈ T1T ðxÞ6 þT2 ð x Þ 7 . 6 7 .. . 6
7 6 7
6 7 6 7 6 f ðm 2Þh; t 7 4 5 f ð m 1 Þh; t 4
5 f ðmh; tÞ f ðm 1Þh; t |fflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} C1
ð1:76Þ
C2
Each element of the coefficient column vectors C1 and C2 in Equation (1.76) can further be expanded by using TFs with respect to the independent variable t. 3 3 2 FT31 T1ðtÞ þ FT41 T2ðtÞ FT11 T1ðtÞ þ FT21 T2ðtÞ 7 7 6 6 FT32 T1ðtÞ þ FT42 T2ðtÞ FT12 T1ðtÞ þ FT22 T2ðtÞ 7 7 6 6 7 7 6 6 .. .. 7 7 6 6 7 7 6 6 . . T T f ðx; tÞ ≈ T1 ðxÞ6 þT2 ð x Þ 7 7; 6 .. .. 7 7 6 6 . . 7 7 6 6 7 7 6 T 6 T T T 4 F1ðm1Þ T1ðtÞ þ F2ðm1Þ T2ðtÞ 5 4 F3ðm1Þ T1ðtÞ þ F4ðm1Þ T2ðtÞ 5 FT1m T1ðtÞ þ FT2m T2ðtÞ FT3m T1ðtÞ þ FT4m T2ðtÞ |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} 2
C1
C2
ð1:77Þ
where h
i FT1ðiþ1Þ ¼ f ðih; 0Þ f ðih; hÞ f ðih; 2hÞ . . . . . . f ih; ðm 1Þh ; i2½0; m 1; h i
FT2ð jþ1Þ ¼ f ð jh; hÞ f ð jh; 2hÞ f ð jh; 3hÞ . . . f jh; ðm 1Þh f ð jh; mhÞ ; j2½0; m 1; h
i FT3k ¼ f ðkh; 0Þ f ðkh; hÞ f ðkh; 2hÞ . . . f kh; ðm 2Þh f kh; ðm 1Þh ; k2½1; m; h i
FT4l ¼ f ðlh; hÞ f ðlh; 2hÞ f ðlh; 3hÞ . . . f lh; ðm 1Þh f ðlh; mhÞ ; l2½1; m:
Equation (1.77) can be written as f ðx; tÞ ≈ T1T ðxÞ F1 T1ðtÞ þ F2 T2ðtÞ þ T2T ðxÞ F3 T1ðtÞ þ F4 T2ðtÞ ;
ð1:78Þ
where F1 , F2 , F3 and F4 are square matrices of size m m, and are defined as
22
Fractional Order Processes
2
2 2 3 3 3 F21 F31 F41 6 F22 7 6 F32 7 6 F42 7 7 6 6 6 7 7 7 7 6 F23 7 6 F33 7 6 7 7 7 ; F2 ¼ 6 7 ; F3 ¼ 6 7 ; F4 ¼ 6 F43 7: 6 .. 7 6 .. 7 6 .. 7 7 4 . 5 4 . 5 4 . 5 5 F1m m m F2m m m F3m m m F4m m m
F11 6 F12 6 6 F1 ¼ 6 F13 6 .. 4 .
3
2
1.6.4 The TF Estimate of the First-Order Integral of Function f ðtÞ One-fold integration of square integrable function f ðtÞ is ðt 1 0 J t f ðt Þ
¼ f ðtÞdt:
ð1:79Þ
0
Substituting the TF estimate of f ðtÞ in Equation (1.68) leads to ðt
ðt f ðtÞdt ffi
0
"
¼ CT
T
Ðt
0
T10 ðtÞdt; . . .
"
T
ðt T1m ðtÞdt þ þD
T
0
......;
0
þD
ðt
C T1m ðtÞ þ D T2m ðtÞ dt ¼ C T
Ðt
T20 ðtÞdt; . . . . . . . . . ;
0
0
#T
T1m1 ðtÞdt
0
Ðt
T2m ðtÞdt;
T
Ðt
#T T2m1 ðtÞdt
:
0
ð1:80Þ Integration of the function f ðtÞ is now changed to integration of LHTF set and RHTF set. Because the function f ðtÞ is square integrable, its TF estimate is also square integrable. The graph of T1i ðtÞ versus t and T2i ðtÞ versus t depicted in Figure 1.2 can be expressed mathematically as T1i ðtÞ ¼ uðt ihÞ T2i ðtÞ ¼
t ih t ði þ 1Þh uðt ihÞ þ u t ði þ 1Þh ; h h
ð1:81Þ
ðt ihÞ t ði þ 1Þh uðt ihÞ u t ði þ 1Þh u t ði þ 1Þh : h h ð1:82Þ
23
Mathematical Postulations
We now integrate each component of LHTF set T1m ðtÞ using Equation (1.81) and express the result in terms of LHTF set T1m ðtÞ and RHTF set T2m ðtÞ. ðt ðt t th uðt hÞdt; T10 ðtÞdt ¼ 1 uðtÞdtþ h h 0 0 h t2 t2 h ¼ t uðtÞ t uðt hÞ; 2 2h 2h h h ¼ T20 ðtÞ þ ½ðT11 ðtÞ þ T21 ðtÞÞ þ ðT12 ðtÞ þ T22 ðtÞÞ þ þ ðT1m1 ðtÞ þ T2m1 ðtÞÞ; 2 2 h h ¼ ½ 0 1 1 . . . 1 1 T1m ðtÞ þ ½ 1 1 1 . . . 1 1 T2m ðtÞ: 2 2 ð1:83Þ
ðt
ðt
h T11 ðtÞdt ¼ ½ 0 2
1 …
0
h 1 1 T1m ðtÞ þ ½ 0 1 2
… 1
1
1 T2m ðtÞ:
0
ð1:84Þ .. . .. . .. . .. . ðt
h T1m1 ðtÞdt ¼ ½ 0 2
0 0
… 0
h 0 T1m ðtÞ þ ½ 0 0 2
0
… 0
1 T2m ðtÞ:
0
ð1:85Þ Therefore, the first-order integration of LHTF set T1m ðtÞ is ðt T1m ðtÞdt ffi P1 T1m ðtÞ þ P2 T2m ðtÞ; 2
0 60 6 6 where P1 ¼ h2 6 0 6. 4 .. 0
0
1 0 0 .. . 0
... 1 .. 0 . .. 0 . ... ... 1 1
2 3 1 1 6 17 60 6 .. 7 7 . 7 ; P2 ¼ h2 6 60 7 6 .. 4. 15 0 m m 0
1
1
1
1
0 .. .
1
0
0 ...
... .. . .. . .. . 0
ð1:86Þ 1
3
7 17 7 17 7: 7 15 1 m m
24
Fractional Order Processes
Following the same procedure, the first-order integration of RHTF set T2m ðtÞ using Equation (1.82) is ðt
ðt T2m ðtÞdt ffi P1 T1m ðtÞ þ P2 T2m ðtÞ ¼ T1m ðtÞdt:
0
ð1:87Þ
0
Equation (1.79) becomes ðt
f ðtÞdt ffi CT þ DT P1 T1m ðtÞ þ P2 T2m ðtÞ :
ð1:88Þ
0
Here P1 and P2 are complement to each other as far as P is considered. This complementary pair is acting as a first-order integral in the TF domain. 1.6.5 The TF Estimate of Riemann-Liouville Fractional-Order Integral of f ðtÞ The Riemann-Liouville fractional-order integral of the function f ðtÞ is
α 0 J t f ðt Þ
ðt 1 ¼ ðt τÞ α 1 f ðτÞdτ: Γð α Þ
ð1:89Þ
0
Using integral convolution property, we rewrite the above equation as
α 0 J t f ðt Þ
ðt
1 1 α 1 t ¼ ðt τÞ α 1 f ðτÞdτ ¼ f ðtÞ : Γð α Þ Γð α Þ
ð1:90Þ
0
Replacing f ðtÞ with its TF estimate, α1 α1 1 α1 T1m ðtÞ T2m ðtÞ T t T t f ðt Þ ¼ C t þD ðαÞ ðαÞ ðαÞ ¼ CT ðJα T1m ðtÞÞ þ DT ðJα T2m ðtÞÞ; α
¼ C ½ J T10 ðtÞ; T
ð1:91Þ α
J T11 ðtÞ; . . . . . . ; J T1m1 ðtÞ
þ D ½ J T20 ðtÞ; Jα T21 ðtÞ; T
α
α
T
. . . . . . ; Jα T2m1 ðtÞ T
25
Mathematical Postulations
Similar to Equation (1.83), we compute the α order Riemann-Liouville fractional integral of T10 ðtÞ and express the result by means of complementary pair of TF sets. α 0 Jt T10 ðtÞ
ðt ðt 1 τ
1 τh α1 dτ; ¼ ðt τ Þ 1 ðt τÞα1 dτ þ ðαÞ h ðαÞ h 0
h
α t αþ1 α t 1 τ tτα ταþ1 1 τ tτ τα þ ; ¼ hα α ðαÞ α hα hðα þ 1Þ 0 ðαÞ hðα þ 1Þ h hα ¼ ½ 0 &1 &2 &m1 T1m ðtÞ ð α þ 2Þ hα ½ & & &m T2m ðtÞ; þ ð α þ 2Þ 1 2 ð1:92Þ
where ςj ¼ j α ð1 þ α jÞ þ ð j 1Þð α þ1Þ , j ¼ 1; 2; …; m 1. α 0 J t T11 ðtÞ ¼
hα hα ½ 0 0 ς1 ς2 . .. ςm2 T1m ðtÞþ ½ 0 ς1 ς2 . .. . .. ςm1 T2m ðtÞ: Γð α þ 2Þ Γð α þ 2Þ ð1:93Þ
.. . .. . .. . .. . α 0 J t T1m1 ðtÞ
¼
hα ½0 Γð α þ 2Þ
0
0
0 ...
0 T1m ðtÞ þ
hα ½0 Γð α þ 2Þ
0
0
0
. . . ς1 T2m ðtÞ: ð1:94Þ
Therefore, the Riemann-Liouville fractional integral of order α of LHTF set T1m ðtÞ is ðt 1 ðt τÞ α 1 T1m ðτÞdτ ¼ P1α T1m ðtÞ þ P2α T2m ðtÞ; Γð α Þ 0
where
ð1:95Þ
26
Fractional Order Processes
2
0 ς1 60 0 6 60 0 hα 6. P1α ¼ 6 Γð α þ 2Þ 6 .. 0 6 4 0 ... 0 ...
ς2 ς1 0 .. . .. . ...
ς3 ς2 ς1 0 .. . 0
... ... ... .. . .. . 0
2 3 ς1 ςm1 60 ςm2 7 6 7 60 ςm3 7 6 hα α 6 .. .. 7 7; P2 ¼ Γð α þ 2Þ 6 . 7 6 . 7 6 40 ς1 5 0 0
ς2 ς1 0 0 .. .
...
ς3 ς2 ς1 .. . .. . ...
ς4 ς3 ς2 ς1 .. . 0
3 . . . ςm . . . ςm1 7 7 . . . ςm2 7 7 .. 7: .. . 7 . 7 7 .. . ς2 5 0 ς1
Following the same procedure as we applied for LHTF set, the RiemannLiouville fractional-order integral of RHTF set T2m ðtÞ using Equation (1.82) is derived as. α 0 J t T2m ðtÞ
ðt 1 ðt τ Þ α 1 T2m ðτÞdτ ¼ P3α T1m ðtÞ þ P4α T2m ðtÞ; ¼ Γð α Þ
ð1:96Þ
0
where 2
0 60 6 60 hα 6. P3α ¼ 6 Γð α þ 2Þ 6 .. 6 40 0
ξ1 0 0 0 .. .
...
ξ2 ξ3 ξ1 ξ2 0 ξ1 .. . 0 .. .. . . ... 0
... ... ... .. . .. . 0
2 3 ξ1 ξ m1 60 ξ m2 7 6 7 60 ξ m3 7 6 hα 6 .. .. 7 7; P4α ¼ Γð α þ 2Þ 6 . 7 6 . 7 6 40 ξ1 5 0 0
ξ2 ξ1 0 0 .. .
...
ξ3 ξ2 ξ1 .. . .. . ...
ξ4 ξ3 ξ2 ξ1 .. . 0
... ... ... .. . .. . 0
ξm
3
ξ m1 7 7 ξ m2 7 7 .. 7; . 7 7 7 ξ2 5 ξ1
ξ j ¼ j α þ1 ð j þ α Þð j 1Þ α :
From Equations (1.89), (1.95) and (1.96), eα 0 J t f ðtÞ ¼
ðt 1 ðt τÞ α 1ef ðτÞdτ ffi CT P1α þ DT P3α T1m ðtÞ þ CT P2α þ DT P4α T2m ðtÞ: Γð α Þ 0
ð1:97Þ
For the special case of α ¼ 1, P1α ¼ P3α ¼ P1 ; P2α ¼ P4α ¼ P2 :
ð1:98Þ
So the TF estimate of fractional-order integral will be reduced to the TF estimate of the first-order integral when α ¼ 1. 1.6.6 Error Analysis Let us denote the TF estimate of function f ðtÞ as
27
Mathematical Postulations
ef ðtÞ ¼ CT T1m ðtÞ þ DT T2m ðtÞ:
ð1:99Þ
We replace f ðtÞ with ef ðtÞ in Equation (1.89) and we call the resulting integral the mth approximate of the α order Riemann-Liouville fractional integral of f ðtÞ. eα 0 J t f ðt Þ ¼
ðt ðt
1 1 α 1e ðt τÞ f ðτ Þdτ ¼ ðt τÞ α 1 CT T1m ðτÞ þ DT T2m ðτÞ dτ: Γð α Þ Γð α Þ 0
0
ð1:100Þ The absolute error between the exact fractional integral 0 J αt f ðtÞ and the mth approximate 0eJ αt f ðtÞ is εm ¼ j0 J αt f ðtÞ 0eJ αt f ðtÞj:
ð1:101Þ
Theorem 1.6.1: If the function f ðtÞ is represented by a complementary pair of LHTF and RHTF sets, then 00 jf ðihÞj þ O m13 ; t2½ih; ði þ 1ÞhÞ; i ¼ 0; 1; 2; …; m 1: 1 2þ α 00 (ii) εm ≤ 2mMT 2 Γð α þ1Þ þ O m3 ; t2½ih; ði þ 1ÞhÞ; jf ðihÞj ≤ M, 8i, M is finite positive value. (i) jf ðtÞ ef ðtÞj ≤
T2 2m2
Proof: (i) From Equations (1.65) to (1.66), we can approximate f ðtÞ in the ith interval as
t ih ef ðtÞ ¼ f ðihÞT1i ðtÞ þ f ði þ 1Þh T2i ðtÞ ¼ f ðihÞ 1 t ih þ f ði þ 1Þh ; h h
0 1
t ih f ði þ 1Þh f ðihÞ t ih Aðt ihÞ; ¼ f ðihÞ þ f ði þ 1Þh f ðihÞ ¼ f ðihÞ þ @ h h h 0
¼ f ðihÞ þ f ðihÞðt ihÞ; h ! 0: ð1:102Þ
Expanding the exact function f ðtÞ by Taylor series with the center ih as 0
f ðtÞ ¼ f ðihÞ þ ðt ihÞf ðihÞ þ
∞ X ðt ihÞ2 00 ðt ihÞk ðkÞ f ðihÞ þ f ðihÞ: 2 k! k¼3
ð1:103Þ
28
Fractional Order Processes
From Equations (1.102) and (1.103), the absolute error between the function and its TF estimate can be determined as jf ðtÞ ef ðtÞj ¼
ðt ihÞ2 00 jf ðihÞj þ Oðt ihÞ3 : 2
ð1:104Þ
Because ðt ihÞ < h and mh ¼ T, the above equation becomes jf ðtÞ ef ðtÞj ¼
T2 00 1 : jf ð ih Þj þ O 2m2 m3
ð1:105Þ
(ii) The absolute error between the exact fractional integral 0 J αt f ðtÞ and the mth approximate 0eJ αt f ðtÞ is ðt 1 εm ¼ 0 Jtα f ðtÞ 0~Jtα f ðtÞ ¼ ðt τÞα1 f ðτÞ ~f ðτÞdτ; ðαÞ 0 2 3 ðrþ1 ðt ð Þh i1 X 1 6 7 α1 α1 ~ ~ ¼ ðt τÞ f ðτÞ f ðτÞdτ þ ðt τÞ f ðτÞ f ðτÞdτ5; 4 ðαÞ r¼0 ih
rh
2 3 ðrþ1 2 2 ð Þh ðt i1 00 1 6X T 00 1 T 1 7 α1 α1 f ðihÞ þ O ðt τÞ f ðihÞ þ O dτ þ ðt τÞ dτ5; 4 ðαÞ r¼0 m3 m3 2m2 2m2 ih
rh
2 3 ðrþ1 X ð Þh ðt i1 1 T 00 1 6 7 α1 α1 f ðihÞ þ O ðt τÞ dτ þ ðt τÞ dτ5; 4 ðαÞ 2m2 m3 r¼0
2
tα T2 00 1 f ðihÞ þ O : ðα þ 1Þ 2m2 m3
rh
ih
ð1:106Þ
We now consider the following assumption. 00
Maxjf ðihÞj ≤ M; 8i; i ¼ 0; 1; 2; …; m 1;
ð1:107Þ
where M is finite positive value. From Equations (1.106) and (1.107), the absolute error between 0 J αt f ðtÞ and 0eJ αt f ðtÞ can be estimated as εm ≤
MT2þ α 1 : þ O 2 2m Γð α þ 1Þ m3
ð1:108Þ
To confirm whether the maximal absolute error caused by TFs will be smaller than the theoretical upper bound derived in Equation (1.108), we consider the function f ðtÞ ¼ t in the interval ½0; 1 divided into five equal subintervals (m ¼ 5).
29
Mathematical Postulations
TABLE 1.1 Absolute errors using TFs and BPFs. t
j0 J αt f ðtÞ 0eJ αt f ðtÞj
α j0 J αt f ðtÞ BPF 0 J t f ðtÞj
0
0
0.0336417669602688
0.2
0
0.0615115192345456
0.4
0
0.0796544626999182
0.6
0
0.0943263867874177
0.8
0
0.106992253349687
1
0
0.118304589557845
The exact fractional integral of function f ðtÞ is α 0 Jt t
¼ ½0 0:067283533920 0:19030657238 0:34961549778 0:53826827136 0:75225277806:
ð1:109Þ
Using Equation (1.97) with α ¼ 0:5, T ¼ 1 and h ¼ 0:2, the TF estimate of J α t is obtained as ~α 0 J t ¼½0 0:067283533920 0:19030657238 0:34961549778 0:53826827136T1m ðtÞ t
þ ½ 0:067283533920 0:19030657238 0:34961549778 0:53826827136 0:75225277806T2m ðtÞ:
ð1:110Þ
In the BPF domain, J α t is approximated as BPF α 0 Jt t
¼ ½0:0336417669 0:128795053 0:2699610350 0:4439418845 0:6452605247 0:870557367621ψm ðtÞ:
ð1:111Þ
Table 1.1 presents the absolute errors given by TF domain analysis and BPF domain analysis. The piecewise linear nature of TFs made them capable of estimating the fractional integral accurately even with small value of m. Therefore, the TF estimate of Riemann-Liouville fractional integral is effective. 1.6.7 MATLAB® Code for Generalized Triangular Function Operational Matrices Programs 1.1 and 1.2 create the generalized triangular function operational matrices for integral of arbitrary order.
30
Fractional Order Processes
Program 1.1 function [P1alph,P2alph]=TOF1(t0,T,m,alpha) % t0 is left end point of the interval [a, b], T is the right end point of the interval, m is the number of subintervals, alpha is the order of integration. h=(T-t0)/(m+0);t=[t0:h:T];P1alph=zeros(length(t)-1, length(t)-1); P2alph=zeros(length(t)-1,length(t)-1);geta1=zeros(1, length(t)); for j=0:1:length(t)-1 if j>0 geta2(j+0)=(j^alpha)*(1+alpha-j)+(j-1)^(alpha+1); end if j>0 geta1(j-0)=(alpha-j)*((j+1)^alpha)+j^(alpha+1); end end geta3=geta2(2:end); for ii=1:1:length(t)-1 for kk=1:1:length(t)-1 if ii==kk P2alph(ii,ii)=geta2(1); end if kk>ii P1alph(ii,kk)=geta2(kk-ii);P2alph(ii,kk)=geta3 (kk-ii); end end end P1alph=((h^alpha)/gamma(alpha+2))*P1alph; P2alph=((h^alpha)/gamma(alpha+2))*P2alph; if alpha==0 P1alph=1;P2alph=1; end end Program 1.2 function [P3alph,P4alph]=TOF12(t0,T,m,alpha) h=(T-t0)/(m);t=[t0:h:T];P1alph=zeros(length(t)-1,length (t)-1); P2alph=zeros(length(t)-1,length(t)-1);geta1=zeros(1, length(t)); for j=0:1:length(t)+0 if j>0 geta2(j+0)=(j^(alpha+1)-(j+alpha)*(j-1)^(alpha));
Mathematical Postulations
31
end if j>0 geta1(j+1)=((j+2)^(alpha+1))-(2+alpha+j)*(j+1) ^alpha; end end geta1(1)=1-(2^alpha)*(1-alpha);geta3=geta2(2:end); for ii=1:1:length(t)-1 for kk=1:1:length(t)-1 if ii==kk P2alph(ii,ii)=geta2(1); end if kk>ii P1alph(ii,kk)=geta2(kk-ii); P2alph(ii,kk)=geta3(kk-ii); end end end P1alph=((h^alpha)/gamma(alpha+2))*P1alph; P2alph=((h^alpha)/gamma(alpha+2))*P2alph; if alpha==0 P1alph=1;P2alph=1; end P3alph= P1alph;P4alph= P2alph; end
1.7 Triangular Strip Operational Matrices for Classical and Fractional Derivatives In this section, we build triangular strip operational matrices from the discretized forms of classical derivative and fractional-order derivative [13]. 1.7.1 Operational Matrix for Classical Derivative Let us first derive triangular strip operational matrix for nth order derivative (n is a positive integer) of casual function f ðtÞ, which is continuous in the interval ½0; T. We start with the first-order derivative of f ðtÞ. Generate N equidistant nodes in the interval ½0; T with step size h, that is, ti ¼ t0 þ ih, i ¼ 0; 1; 2; ……; N. 0 To find out the numerical solution of f ðtÞ, the following first-order backward difference formula can be used.
32
Fractional Order Processes
0
f ðti Þ ffi
1 1 rf ðti Þ ¼ ðfi fi1 Þ; i ¼ 0; 1; 2; ……; N: h h
ð1:112Þ
From the above formula, N expressions can be written as 1 1 rf ðt0 Þ ¼ ðf0 f1 Þ at node 1; h h 1 1 rf ðt1 Þ ¼ ðf1 f0 Þ at node 2; h h .. . 1 1 rf ðtN Þ ¼ ðfN fN1 Þ at node N: h h
ð1:113Þ
The system of equations in Equation (1.113) can be shown in the vectormatrix form below. 2
3 h1 f ðt0 Þ
3 2 1 0 0 0 . . . 0 2 f0 3 6 1 7 6 h r f ðt1 Þ 7 6 1 1 0 ... ... 07 6 6 f1 7
7 6 6 1 7 7 .. 7 76 6 6 h r f ðt2 Þ 7 f2 7 7 6 0 ... . 6 1 6 0 1 1 6 7 6 7 .. .. .. 7 6 7 ¼ 6 .. 6 .. 7 ; .. 7 6 7 6 h 0 . . 0 .76 . 7 6 . .
7 6 7 74 6 .. 6 h1 r f ðt Þ 7 5 4 0 6 7 0 0 . 1 0 5 fN1 N1 4
5 fN 0 0 . . . 0 1 1 h1 r f ðtN Þ |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} |fflfflfflfflffl{zfflfflfflfflffl} FN |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} B1N
ð1:114Þ
Y
where FN is the vector of function values obtained at equidistant nodes, B1N is the lower triangular strip operational matrix, and Y is the vector of approximated values of f 0 ðtÞ. Likewise, we can get the following triangular strip operational matrix for nth order derivative of f ðtÞ. 2
w0 6 w1 6 6 w2 6 . 6 . 6 . 1 6 BnN ¼ 6 h 6 wn 6 6 6 0 6 . 4 .. 0
0 w0 w1
0 0 w0
0 0 0
w2 .. . .. .
w1 .. . .. . .. .
w0 .. . .. . .. .
0
wn
0 ...
... ... ... ... ... ... 0 .. . .. . .. .
...
... 0 .. . .. . w2
0 0 0 .. . 0 0 .. . w1
0 0 0 .. .
3
7 7 7 7 7 7 7 7; 0 7 7 7 0 7 7 0 5 w0
ð1:115Þ
33
Mathematical Postulations n where wj ¼ ð1Þ , j ¼ 0; 1; 2; …::; N. j j
1.7.2 Operational Matrix for Fractional-Order Derivative Let us recall the definition of fractional-order derivative in the GrunwaldLetnikov sense: GL α 0 D t f ðt Þ ffi
k
X 1 α i α α r f ðt k Þ ¼ h ð1Þ f ðt ihÞ ; i hα i¼0
ð1:116Þ
where k ¼ 0; 1; 2; …; N. There are such N þ 1 equations. Each equation computes the approximate value of f α ðtk Þ at node tk and these equations are simultaneously written in the vector-matrix form given in the next equation. 2
3
2 3 f0 6 7 6 7 6 7 6 f1 7 7 6 7 6 76 6 7 6 7 6 f2 7 1 6 7 6 7 76 6 . 7; 6 7 ¼ α6 ... ... 7 6 7 7 6 .. 7 h 6 7 6 7 6 α 76 4 fN1 5 4 h α r α f ðtN1 Þ 5 6 wN1 w2α w1α w0α 0 7 4 5 .. fN h α r α f ðtN Þ α wNα . w2α w1α w0α |fflfflfflfflffl{zfflfflfflffl wN1 ffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} FN Y 2
α
α
h r f ðt 0 Þ h α r α f ðt1 Þ h α r α f ðt2 Þ .. .
3
w0α w1α w2α .. .
0 w0α w1α .. . .. .
0 0 w0α .. .
0 0 0 .. .
... ... ...
0 0 0
BNα
ð1:117Þ
α where wkα ¼ ð1Þk , k ¼ 0; 1; 2; …::; N. k Here, FN is the vector of function values obtained at equidistant nodes, BNα is the lower triangular strip operational matrix for fractional derivatives, and Y is the vector of approximated values of GL0 D αt f ðtÞ. 1.7.3 MATLAB Code for Triangular Strip Operational Matrices The following program generates triangular strip operational matrix for derivative of arbitrary order. Program 1.3 function B = ban(alpha,N,h) % alpha – order of differentiation (real, not necessarily integer) % N – size of the resulting matrix B (N x N) % h – step of discretization; default is h=1 B = zeros(N,N);
34
Fractional Order Processes
if nargin <= 1 || nargin > 3 error(‘BAN: Wrong number of input parameters’) else bc=fliplr(bcrecur(alpha,N-1)); for k=1:N B(k,1:k)=bc((N-k+1):N); end endif nargin == 3 B=1*B; end end % bcrecur.m is called in the main program ‘ban.m’ function y=bcrecur(a, n) y=cumprod([1, 1 – ((a+1) ./ (1:n))]); end
References [1] K. B. Oldham, J. Spanier (1974). The Fractional Calculus: Theory and Applications of Differentiation and Integration to Arbitrary Order. New York: Dover Publications. [2] M. G. Mittag-Leffler (1903). Sur la nouvelle fonction Eα(x). Comptes Rendus Acad. Sci. Paris., vol. 137, pp. 554–558. [3] R. Gorenflo, A. A. Kilbas, S. V. Rogosin (1998). On the generalized Mittag-Leffler type functions. Integral Transform. Spec. Funct., vol. 7, pp. 215–224. [4] I. Podlubny (1999). Fractional Differential Equations. New York: Academic Press. [5] C. Li, W. Deng (2007). Remarks on fractional derivatives. Appl. Math. Comput., vol. 187, pp. 777–784. [6] M. Caputo (1967). Linear models of dissipation whose Q is almost frequency – Part II. J. Roy. Austral. Soc., vol. 13, pp. 529–539. [7] S. Das (2011). Functional Fractional Calculus. Berlin: Springer-Verlag. [8] N. Heymans, I. Podlubny (2005). Physical interpretation of initial conditions for fractional differential equations with Riemann-Liouville fractional derivatives. Rheol. Acta, vol. 45, pp. 765–771. [9] C. A. Monje, Y. Q. Chen, B. M. Vinagre, D. Xue, V. Feliu (2010). Fractional-Order Systems and Controls: Fundamentals and Applications. London: Springer-Verlag. [10] A. Deb, A. Dasgupta, G. Sarkar (2006). A new set of orthogonal functions and its application to the analysis of dynamic systems. J. Franklin Inst., vol. 343, pp. 1–26. [11] A. Deb, G. Sarkar, A. Sengupta (2007). Triangular Orthogonal Functions for the Analysis of Continuous Time Systems. Gurgaeon, India: Elsevier. [12] Y. Li, N. Sun (2011). Numerical solution of fractional differential equations using the generalized block pulse operational matrix. Comput. Math. Appl., vol. 62, pp. 1046–1054. [13] I. Podlubny (2000). Matrix approach to discrete fractional calculus. Fract. Calc. Appl. Anal., vol. 3, pp. 359–386.
Mathematical Postulations K. B. Oldham , J. Spanier (1974). The Fractional Calculus: Theory and Applications of Differentiation and Integration to Arbitrary Order . New York: Dover Publications. M. G. Mittag-Leffler (1903). Sur la nouvelle fonction E(x). Comptes Rendus Acad. Sci. Paris. , vol. 137, pp. 554558. R. Goreno , A. A. Kilbas , S. V. Rogosin (1998). On the generalized Mittag-Leffler type functions. Integral Transform. Spec. Funct., vol. 7, pp. 215224. I. Podlubny (1999). Fractional Differential Equations . New York: Academic Press. C. Li , W. Deng (2007). Remarks on fractional derivatives. Appl. Math. Comput., vol. 187, pp. 777784. M. Caputo (1967). Linear models of dissipation whose Q is almost frequency Part II. J. Roy. Austral. Soc., vol. 13, pp. 529539. S. Das (2011). Functional Fractional Calculus . Berlin: Springer-Verlag. N. Heymans , I. Podlubny (2005). Physical interpretation of initial conditions for fractional differential equations with Riemann-Liouville fractional derivatives. Rheol. Acta , vol. 45, pp. 765771. C. A. Monje , Y. Q. Chen , B. M. Vinagre , D. Xue , V. Feliu (2010). Fractional-Order Systems and Controls: Fundamentals and Applications . London: Springer-Verlag. A. Deb , A. Dasgupta , G. Sarkar (2006). A new set of orthogonal functions and its application to the analysis of dynamic systems. J. Franklin Inst., vol. 343, pp. 126. A. Deb , G. Sarkar , A. Sengupta (2007). Triangular Orthogonal Functions for the Analysis of Continuous Time Systems . Gurgaeon, India: Elsevier. Y. Li , N. Sun (2011). Numerical solution of fractional differential equations using the generalized block pulse operational matrix. Comput. Math. Appl., vol. 62, pp. 10461054. I. Podlubny (2000). Matrix approach to discrete fractional calculus. Fract. Calc. Appl. Anal., vol. 3, pp. 359386.
Numerical Method for Simulation of Physical Processes Represented by Weakly Singular Fredholm, Volterra, and Volterra-Fredholm Integral Equations I. Fredholm (1900). On a new method for solving the Dirichlet problem (Sur une nouvelle mthode pour la rsolution du problme de Dirichlet) (in French). Stockh. Ofv., vol. 57, pp. 3946. E.I. Fredholm (1903). Sur une classe dquations fonctionnelles. Acta Math., vol. 27, pp. 365390. A.D. Polyanin , A.V. Manzhirov (2008). Handbook of Integral Equations . Boca Raton, FL: Chapman & Hall/CRC, Taylor & Francis Group. R. Estrada , R. Kanwal (2000). Singular Integral Equations . Berlin: Birkhauser. H. Adibi , P. Assari (2010). On the numerical solution of weakly singular Fredholm integral equations of the second kind using Legendre wavelets. J. Vibr. Contr. , vol. 17(5), pp. 689698. U. Lepik , E. Tamme (2007). Solution of nonlinear Fredholm integral equations via the Haar wavelet method. Proc. Estonian Acad. Sci. Phys. Math., vol. 56, pp. 1727. A. Pedas , G. Vainikko (1997). Superconvergence of piecewise polynomial collocations for nonlinear weakly singular integral equations. J. Integral. Equ. Appl. , vol. 9(4), pp. 379406. H. Kaneko , Y. Xu (1991). Numerical solutions for weakly singular Fredholm integral equations of the second kind. Appl. Numer. Math. , vol. 7, pp. 167177. K. Maleknejad , A. Ostadi (2017). Using Sinc-collocation method for solving weakly singular Fredholm integral equations of the first kind. Appl. Anal. , vol. 96(4), pp. 702712. S.S. Allaei , T. Diogo , M. Rebelo (2017). Analytical and computational methods for a class of nonlinear singular integral equations. Appl. Numer. Math ., vol. 114, pp. 217. J.B. Keller , W.E. Olmstead (1971-72). Temperature of a nonlinearly radiating semi-infinite solid. Quart. Appl. Math., vol. 29, pp. 559566. F.D. Hoog , R. Weis (1974). High order methods for a class of Volterra integral equations with weakly singular kernels. SIAM J. Numer. Anal., vol. 11(6), pp. 11661180. A.M. Wazwaz (1996). A reliable technique for solving the weakly singular second-kind Volterra type integral equations. Appl. Math. Comput. , vol. 80, pp. 287299. C.C. Lin (1959). Hydrodynamics of liquid helium II. Phys. Hev. Letters , vol. 2, pp. 245246. A.P. Orsi (1996). Product integration for Volterra integral equations of the second kind with weakly singular kernels. Math. Comp., vol. 65, pp. 12011212.
N. Levinson (1960). A nonlinear Volterra equation arising in the theory of superfluidity. J. Math. Anal. Appl., vol. 1, pp. 111. J.M. Lighthill (1950). Contributions to the theory of the heat transfer through a laminary boundary layer. Proc. Roy. Soc., vol. 202, pp. 359377. T. Diogo , P. Lima , M. Rebelo (2006). Numerical solution of a nonlinear Abel type Volterra integral equation. Commun. Pure Appl. Anal., vol. 5, pp. 277288. 71 M.A.F. Araghi , H.D. Kasmaei (2008). Numerical solution of the second kind singular Volterra integral equations by modified Navot-Simpsons quadrature. Int. J. Open Problems Compt. Math., vol. 1(3), pp. 201212. L. Tao , H. Yong (2006). Extrapolation method for solving weakly singular nonlinear Volterra integral equations of the second kind. J. Anal and App. Math., vol. 324, pp. 225237.
Numerical Method for Simulation of Physical Processes Modeled by Abels Integral Equations R. Goreno , S. Vessella (1991). Abel Integral Equations: Analysis and Applications . Berlin: Springer. H. Brunner (1997). 1896-1996: One hundred years of Volterra integral equation of the rst kind. Appl. Numer. Math. , vol. 24, pp. 8393. F.G. Tricomi (1957). Integral Equations . New York: Interscience. D.A. Murio , D.G. Hinestroza , C.W. Mejia (1992). New stable numerical inversion of Abels integral equation. Comput. Math. Appl. , vol. 11, pp. 311. M. Deutsch , I. Beniaminy (1982). Derivative-free inversion of Abels integral equation. Appl. Phys. Lett. , vol. 41, pp. 2728. M. Bocher (1914). An introduction to the study of integral equations, 2nd edn . London: Cambridge University Press. H. Malinowski , R. Smarzewski (1978). A numerical method for solving the Abel integral equation. Appl. Math. , vol. 16(2), pp. 275281. C. Lubich (1983). Runge-Kutta theory for Volterra and Abel integral equations of the second kind. Math. Comput. , vol. 41(163), pp. 87102. 105 C. Lubich (1985). Fractional linear multistep methods for Abel-Volterra integral equations of the second kind. Math. Comput. , vol. 45(172), pp. 463469. R. Gorenflo , Y. Luchko (1997). Operational method for solving generalized Abel integral equation of the second kind. Integr. Transf. Spec. F. , vol. 5, pp. 4758. V. Sizikov , D. Sidorov (2016). Generalized quadrature for solving singular integral equations of Abel type in application to infrared tomography. Appl. Numer. Math. , vol. 106, pp. 6978. R.K. Pandey , O.P. Singh , V.K. Singh (2009). Efficient algorithms to solve singular integral equations of Abel type. Comput. Math. Appl. , vol. 57, pp. 664676. S. Kumar , A. Kumar , D. Kumar , J. Singh , A. Singh (2015). Analytical solution of Abel integral equation arising in astrophysics via Laplace transform. J. Egyptian Math. Soc. , vol. 23, pp. 102107. M.S. Mohamed , K.A. Gepreel , F.A. Al-Malki , M. Al-Humyani (2015). Approximate solutions of the generalized Abels integral equations using the extension Khans Homotopy analysis transformation method. J. Appl. Math. , vol. 2015, Article ID 357861, 9 pages. M. Khan , M.A. Gondal (2012). A reliable treatment of Abels second kind singular integral equations. Appl. Math. Lett. , vol. 25, pp. 16661670. M. Gulsu , Y. Ozturk , M. Sezer (2011). On the solution of the Abel equation of the second kind by the shifted Chebyshev polynomials. Appl. Math. Comput. , vol. 217, pp. 48274833. R.K. Pandey , S. Suman , K.K. Singh , O.P. Singh (2014). An approximate method for Abel inversion using Chebyshev polynomials. Appl. Math. Comput. , vol. 237, pp. 120132. S. Dixit , O.P. Singh , S. Kumar (2012). A stable numerical inversion of generalized Abels integral equation. Appl. Numer. Math. , vol. 62, pp. 567579. M.P. Tripathi , R.K. Pandey , V.K. Baranwal , O.P. Singh (2013). Generalized Abel inversion using extended hat functions operational matrix. Int. J. Anal. , vol. 2013, Article ID 652541, 12 pages. M.N. Sahlan , H.R. Marasi , F. Ghahramani (2015). Block-pulse functions approach to numerical solution of Abels integral equation. Cogent Math. , vol. 2(1). DOI: 10.1080/23311835.2015.1047111. S.A. Youse (2006). Numerical solution of Abels integral equation by using Legendre wavelets. Appl. Math. Comput. , vol. 175, pp. 574580.
K. Maleknejad , M. Nosrati , E. Najafi (2012). Wavelet Galerkin method for solving singular integral equations. Comput. Math. Appl. , vol. 31(2), pp. 373390. X. Li , T. Tang (2012). Convergence analysis of Jacobi spectral collocation methods for AbelVolterra integral equations of second kind. Front. Math. China. , vol. 7(1), pp. 6983. M.A. Fariborzi-Araghi , G. Kazemi-Gelian (2014). The combined Sinc-Taylor expansion method to solve Abels integral equation. Theory. Approx. Appl. , vol. 10(1), pp. 2739. E. Babolian , H.R. Marzban , M. Salmani (2008). Using triangular orthogonal functions for solving Fredholm integral equations of the second kind. Appl. Math. Comput. , vol. 201, pp. 452463. 106 K. Maleknejad , Z. JafariBehbahani (2012). Applications of two-dimensional triangular functions for solving nonlinear class of mixed VolterraFredholm integral equations. Math. Comput. Model. , vol. 55, pp. 18331843. K. Maleknejad , H. Almasieh , M. Roodaki (2010). Triangular functions (TF) method for the solution of nonlinear VolterraFredholm integral equations. Commun. Nonlinear. Sci. , vol. 15, pp. 32933298. E. Babolian , K. Maleknejad , M. Roodaki , H. Almasieh (2010). Two-dimensional triangular functions and their applications to nonlinear 2D VolterraFredholm integral equations. Comput. Math. Appl. , vol. 60, pp. 17111722. F. Mirzaee , S. Piroozfar (2010). Numerical solution of the linear two-dimensional Fredholm integral equations of the second kind via two-dimensional triangular orthogonal functions. J. King. Saud. Univ. , vol. 22, pp. 185193. F. Mirzaee , E. Hadadiyan (2016). Three-dimensional triangular functions and their applications for solving nonlinear mixed VolterraFredholm integral equations. Alexandria Eng. J. , vol. 53(3), pp. 29432952. M. Khodabin , K. Maleknejad , F.H. Shekarabi (2013). Application of triangular functions to numerical solution of stochastic Volterra integral equations. IAENG Int. J. Appl. Math. , vol. 43(1), pp. 19. Z. Sadati (2013). Numerical implementation of triangular functions for solving a stochastic nonlinear Volterra-Fredholm integral equation. IAENG Int. J. Appl. Math. , vol. 45(2), pp. 102107. F. Mirzaee , M.K. Yari , E. Hadadiyan (2015). Numerical solution of two-dimensional fuzzy Fredholm integral equations of the second kind using triangular functions. Beni-Suef Univ. J. Basic Appl. Sci. , vol. 4, pp. 109118. H. Saeedi , G.N. Chuev (2015). Triangular functions for operational matrix of nonlinear fractional Volterra integral equations. J. Appl. Math. Comput. , vol. 49, pp. 213232. S. Noeiaghdam , E. Zarei , H.B. Kelishami (2016). Homotopy analysis transform method for solving Abels integral equations of the first kind. Aim Shams Eng. J. , vol. 7, pp. 483495. H. Derili , S. Sohrabi (2008). Numerical solution of singular integral equations using orthogonal functions. Math. Sci. , vol. 2(3), pp. 261272. S. Suman , K.K. Singh , R.K. Pandey (2014). Approximate solution of integral equation using Bernstein polynomial multiwavelets. In: M Pant , K Deep , A Nagar , J Bansal (eds) Proceedings of the Third International Conference on Soft Computing for Problem Solving. Advances in Intelligent Systems and Computing , vol. 259, New Delhi, India: Springer, pp. 486496. M.M. Rahman (2013). Numerical solutions of Volterra integral equations using Galerkin method with Hermite Polynomials. In Proceedings of the 2013 International Conference on Applied Mathematics and Computational Methods in Engineering , pp. 276281. A. Shahsavaran (2011). Numerical approach to solve second kind Volterra integral equation of Abel type using Block-Pulse functions and Taylor expansion by collocation method. Appl. Math. Sci. , vol. 5(14), pp. 685696. R.A. Chiquet , P. Jara , K.W. Zito (2015). Approximate solutions of Abels equation using rational inversion of the Laplace transform. Neur. Parallel Sci. Comput. , vol. 23, pp. 169178. C. Yang (2014). An efficient numerical method for solving Abel integral equation. Appl. Math. Comput. , vol. 227, pp. 656661. 107 Z. Avazzadeh , B. Shafiee , G.B. Loghmani (2011). Fractional calculus for solving Abels integral equations using Chebyshev polynomials. Appl. Math. Sci. , vol. 5(45), pp. 22072216. A. Saadatmandi , M. Dehghan (2008). A collocation method for solving Abels integral equations of first and second kinds. Z. Naturforsch. , vol. 63(12), pp. 752756. S. Sohrabi (2011). Comparison Chebyshev wavelets method with BPFs method for solving Abels integral equations. Ain Shams Eng. J. , vol. 2, pp. 249253. L.K. Bieniasz (2008). Cyclic Voltammetric Current Functions Determined with a Prescribed Accuracy by the Adaptive Huber Method for Abel Integral Equations. Anal Chem , vol. 80, pp. 96599665.
L.K. Bieniasz (2012). Automatic simulation of electrochemical transients by the adaptive Huber method for Volterra integral equations involving Kernel terms exp[(t )] erex {[(t )] 1/2} and exp [(t )] daw {[(t )] 1/2}. J. Math. Chem. , vol. 50, pp. 765781.
Numerical Method for Simulation of Physical Processes Described by Fractional-Order Integro-Differential Equations J. Hou , B. Qin , C. Yang (2012). Numerical solution of nonlinear Fredholm integrodifferential equations of fractional order by using hybrid functions and the collocation method. J. Appl. Math. , vol. 2012, Article ID 687030, 11 pages. 133 M. Yi , L. Wang , J. Huang (2016). Legendre wavelets method for the numerical solution of fractional integro-differential equations with weakly singular kernel. Appl. Math. Model. , vol. 40, pp. 34223437. Y. Wang , L. Zhu (2016). SCW method for solving the fractional integro-differential equations with a weakly singular kernel. Appl. Math. Comput. , vol. 275, pp. 7280. M.X. Yi , J. Huang (2015). CAS wavelet method for solving the fractional integro-differential equation with a weakly singular kernel. Int. J. Comput. Math., vol. 92(8), 17151728. K. TeBeest (1997). Numerical and analytical solutions of Volterras population model. SIAM Rev., vol. 39, pp. 484493. Y. Alwesabi , A.A. Dahawi , Y.S. Daniel , A.H.M. Murid (2014). Analytical solution of Volterras population model using Variation Iteration Method (VIM). In Proceedings of 1st International Conference of Recent Trends in Information and Communication Technologies . S.T. Mohyud-Din , A. Yildirim , Y. Gulkanat (2010). Analytical solution of Volterras population model. J. King Saud Univ. , vol. 22, pp. 247250. V.K. Singh , E.B. Postnikov (2013). Operational matrix approach for solution of integrodifferential equations arising in theory of anomalous relaxation processes in vicinity of singular point. Appl. Math. Model. , vol. 37, pp. 66096616.
Numerical Method for Simulation of Physical Processes Represented by Stiff and Nonstiff Fractional-Order Differential Equations, and Differential-Algebraic Equations K.B. Oldham , J. Spanier (1974). The Fractional Calculus: Theory and Applications of Differentiation and Integration to Arbitrary Order . New York: Dover Publications. T.M. Atanackovic , S. Pilipovic , B. Stankovic , D. Zorica (2014). Fractional Calculus with Applications in Mechanics: Vibrations and Diffusion Processes . London: Wiley. R. Herrmann (2011). Fractional Calculus: An Introduction for Physicists . Singapore: World Scientic Publishing Co. Pte. Ltd. G.A. Losa , D. Merlini , T.F. Nonnenmacher , E.R. Weibel (2005). Fractals in Biology and Medicine . Basel: Birkhauser. F. Mainardi (2010). Fractional Calculus and Waves in Linear Viscoelasticity . London: Imperial College Press. A.K. Golmankhneh , X.J. Yang , D. Balean (2015). Einstein field equations within local fractional calculus. Rom. J. Phys., vol. 60, pp. 2231. M. Biyajima , T. Mizoguchi , N. Suzuki (2015). A new blackbody radiation law based on fractional calculus and its application to NASA COBE data. Physica A , vol. 440, pp. 129138. W.M. Ahmad , R. El-Khazali (2007). Fractional-order dynamical models of love. Chaos Soliton. Fract., vol. 33, pp. 13671375. L. Song , S. Xu , J. Yang (2007). Dynamical models of happiness with fractional order. Commun. Nonlinear Sci. Numer. Simulat., vol. 15, pp. 616628. I. Podlubny (1999). Fractional Differential Equations . New York: Academic Press. K. Diethelm (1997). An algorithm for the numerical solution of differential equations of fractional order. Electron. Trans. Numer. Anal., vol. 5, pp. 15. Z.M. Odibat , S. Momani (2008). An algorithm for the numerical solution of differential equations of fractional order. J. Appl. Math. Informatics , vol. 26, pp. 1527. G. Adomian (1994). Solving Frontier Problems of Physics: The Decomposition Method . Boston: Kluwer Academic Publishers.
J.H. He (1999a). Variational iteration method A kind of non-linear analytical technique: Some examples. Int. J. Non-Linear Mech., vol. 34, pp. 699708. J.H. He (1999b). Homotopy perturbation technique. Comput. Method Appl. Mech. Eng., vol. 178, pp. 257262. A. Arigoklu , I. Ozkol (2007). Solution of fractional differential equations by using differential transform method. Chaos Soliton Fract., vol. 34, pp. 14731481. M. Zurigat , S. Momani , A. Alawneh (2010). Analytical approximate solutions of systems of fractional algebraicDifferential equations by homotopy analysis method. Comput. Math. Appl. , vol. 59, pp. 12271235. 189 B. Ibi , M. Bayram (2011). Numerical comparison of methods for solving fractional differential-algebraic equations (FDAEs). Comput. Math. Appl. , vol. 62, pp. 32703278. B. Ibi , M. Bayram , A.G. Agargun (2011). Applications of fractional differential transform method to fractional differential-algebraic equations. Europ. J. Pure Appl. Math. , vol. 4(2), pp. 129141. O.S. Odetunde , O.A. Taiwo (2015). An algorithm for the approximation of fractional differentialalgebraic equations with Caputo-type derivatives. J. Appl. Computat. Math., vol. 4, pp. 242. DOI:10.4172/2168-9679.1000242. H.M. Jaradat , M. Zurigat , S. Al-Shara , Q. Katatbeh (2014). Toward a new algorithm for systems of fractional differential equations. Italian J. Pure Appl. Math. , vol. 32, pp. 579594. X.-L. Ding , Y.-L. Jian (2014). Waveform relaxation method for fractional differential-algebraic equations. Fract. Calc. Appl. Anal., vol. 17(3), pp. 585604. N.J. Ford , J.A. Connolly (2009). Systems-based decomposition schemes for the approximate solution of multi-term fractional differential equations. J. Comput. Appl. Math. , vol. 229, pp. 382391. S.C. Shiralashetti , A.B. Deshi (2016). An efcient Haar wavelet collocation method for the numerical solution of multi-term fractional differential equations. Nonlinear Dyn., vol. 83, pp. 293303. E. Hesameddini , A. Rahimi , E. Asadollahifard (2016). On the convergence of a new reliable algorithm for solving multi-order fractional differential equations. Commun. Nonlinear Sci. Numer. Simulat., vol. 34, pp. 154164. A.E.M. El-Mesiry , A.M.A. El-Sayed , H.A.A. El-Saka (2005). Numerical methods for multi-term fractional (arbitrary) orders differential equations. Appl. Math. Comput. , vol. 160, pp. 683699. M.A. EL-Sayed , A.E.M. EL-Mesiry , H.A.A. EL-Saka (2004). Numerical solution for multi-term fractional (arbitrary) orders differential equations. Comput. Appl. Math. , vol. 23(1), pp. 3354. A.M.A. El-Sayed , M.M. Saleh , E.A.A. Ziada (2010). Analytical and numerical solution of multiterm nonlinear differential equations of arbitrary orders. J. Appl. Math. Comput., vol. 33, pp. 375388. M. Javidi , N. Nyamoradi (2013). A numerical scheme for solving multi-term fractional differential equations. Commun. Frac. Calc., vol. 4(1), pp. 3849. K.D. Kucche , J.J. Trujillo (2017). Theory of system of nonlinear fractional differential equations. Progr. Fract. Differ. Appl., vol. 3(1), pp. 718. W.K. Zahra , M.V. Daele (2017). Discrete spline methods for solving two point fractional BagleyTorvik equation. Appl. Math. Comput. , vol. 296, pp. 4255. S. Yuzbasi (2013). Numerical solution of the BagleyTorvik equation by the Bessel collocation method. Math. Meth. Appl. Sci., vol. 36, pp. 300312. M.U. Rehman , R.A. Khan (2012). A numerical method for solving boundary value problems for fractional differential equations. Appl. Math. Model., vol. 36, pp. 894907. Q. Din , A.A. Elsadany , H. Khalil (2017). Neimark-Sacker Bifurcation and Chaos Control in a Fractional-Order Plant-Herbivore Model. Discrete. Dynam. Nat. Soc. , vol. 2017, Article ID 6312964, 15 pages. 190 W.-C. Chen (2008). Nonlinear dynamics and chaos in a fractional-order nancial system. Chaos Soliton. Fract., vol. 36, pp. 13051314. A.A. Freihat , M. Zurigat , A.H. Handam (2015). The multi-step homotopy analysis method for modied epidemiological model for computer viruses. Afr. Mat., vol. 26(3), pp. 585595. F. Mazzia , F. Iavernaro (2003). Test set for initial value problem solvers, Tech. Rep 40. Department of Mathematics, University of Bari, Italy. H. Shintani (1982). Modified Rosenbrock methods for stiff systems. Hiroshima Math. J., vol. 12, pp. 543558. A. Elsaid (2012). Fractional differential transform method combined with the Adomian polynomials. Appl. Math. Comput., vol. 218, pp. 68996911. E. Schafer (1975). A new approach to explain the high irradiance responses of photomorphogenesis on the basis of phytochrome. J. Math. Biology , vol. 2, pp. 4155.
Numerical Method for Simulation of Fractional Diffusion-Wave Equation W. Deng , C. Li , Q. Guo (2007). Analysis of fractional differential equations with multi-orders. Fractals ., vol. 15, pp. 173182. M. Aslefallah , E. Shivanian (2015). Nonlinear fractional integro-differential reactionDiffusion equation via radial basis functions. Eur. Phys. J. Plus , vol. 47, pp. 19. M.M. Meerschaert , C. Tadjeran (2006). Finite difference approximations for two-sided spacefractional partial differential equations. Appl. Numer. Math., vol. 56, pp. 8090. R. Metler , J. Klafter (2004). The restaurant at the end of random walk: Recent developments in the description of anomalous transport by fractional dynamics. J. Phys. A , vol. 37, pp. 161208. G.M. Zaslavsky (2002). Chaos, fractional kinetics, and anomalous transport. Phys. Rep., vol. 371, pp. 461580. Z. Wang , S. Vong (2014). Compact difference schemes for the modied anomalous fractional sub-diffusion equation and the fractional diffusion-wave equation. J. Comput. Phys., vol. 277, pp. 115.
Identification of Fractional Order Linear and Nonlinear Systems from Experimental or Simulated Data A. Benchellal , T. Poinot , J.C. Trigeassou (2007). Modelling and Identification of Diffusive Systems using Fractional Models. In J. Sabatier , O.P. Agrawal , J.A.T. Machado (eds) Advances in Fractional Calculus . Dordrecht: Springer. A. Benchellal , T. Poinot , J.-C. Trigeassou (2006). Approximation and identification of diffusive interfaces by fractional models. Signal Process. , vol. 86(10), pp. 27122727. J.C. Wang (1987). Realizations of generalized Warburg impedance with RC ladder networks and transmission lines. J. Electrochem. Soc. , vol. 134(8), pp. 19151920. A. Jalloul , K. Jelassi , P. Melchior , J.-C. Trigeassou (2011). Fractional identification of rotor skin effect in induction machines. Int. J. Comput. Sci. , vol. 8(1), pp. 5767. R.L. Bagley , R.A. Calico (1991). Fractional order state equations for the control of viscoelasticallydamped structures. J. Guid. Cont. Dyn. , vol. 14(2), pp. 304311. W.M. Ahmad , R. El-Khazali (2007). Fractional-order dynamical models of love. Chaos Soliton Fract , vol. 33, pp. 13671375. L. Song , S. Xu , J. Yang (2010). Dynamical models of happiness with fractional order. Commun. Nonlinear Sci , vol. 15(3), pp. 616627. A.A. Freihat , M. Zurigat , A.H. Handam (2015). The multi-step homotopy analysis method for modied epidemiological model for computer viruses. Afrika Mat. , vol. 26, pp. 585597. Y. Cho , I. Kim , D. Sheen (2015). A fractional-order model for MINMOD Millennium. Math. Biosci. , vol. 262, pp. 3645. Q. Din , A.A. Elsadany , H. Khalil (2017). Neimark-Sacker bifurcation and chaos control in a fractional-order Plant-Herbivore model. Discrete. Dyn. Nat. Soc. , vol. 2017, Article ID 6312964, 15. H.M. Srivastava , D. Kumar , J. Singh (2017). An efficient analytical technique for fractional model of vibration equation. Appl. Math. Model. , vol. 45, pp. 192204. J.K. Popovic , M.T. Atanackovic , A.S. Pilipovic , M.R. Rapaic , S. Pilipovic , T.M. Atanackovic (2010). A new approach to the compartmental analysis in pharmacokinetics: Fractional time evolution of diclofenac. J. Pharmacokinet. Phar. , vol. 37(2), pp. 119134. A.D. Freed , K. Diethelm (2006). Fractional calculus in biomechanics: A 3D viscoelastic model using regularized fractional derivative kernels with application to the human calcaneal fat pad. Biomech. Model. Mechan. , vol. 5, pp. 203215. A.M. Lopes , J.A.T. Machado , E. Ramalho (2017). On the fractional-order modeling of wine. Eur. Food. Res. Technol. , vol. 243(6), pp. 921929. N. Heymans , I. Podlubny (2005). Physical interpretation of initial conditions for fractional differential equations with Riemann-Liouville fractional derivatives. Rheol. Acta. , vol. 45(5), pp. 765771. A. Dzielinski , D. Sierociuk , G. Sarwas , I. Petras , I. Podlubny , T. Skovranek (2011). Identication of the fractional-order systems: A frequency domain approach. Acta. Montan. Slovaca. , vol. 16(1), pp. 2633. D. Valerio , M. Ortigueira , J. Sa Da Costa (2008). Identifying a transfer function from a frequency response. J. Comput. Nonlinear Dyn. , vol. 3(2), pp. 021207-1021207-7. DOI: 10.1115/1.2833906.
D. Valrio , I. Tejado (2015). Identifying a non-commensurable fractional transfer function from a frequency response. Signal Process. , vol. 107, pp. 254264. 234 P. Nazarian , M. Haeri , M.S. Tavazoei (2010). Identifiability of fractional order systems using input output frequency contents. ISA T. , vol. 49(2), pp. 207214. A. Djouambi , A. Voda , A. Charef (2012). Recursive prediction error identication of fractional order models. Commun. Nonlinear. Sci. , vol. 17(6), pp. 25172524. T. Poinot , J.-C. Trigeassou (2004). Identication of fractional systems using an output-error technique. Nonlinear Dyn. , vol. 38, pp. 133154. A. Maachou , R. Malti , P. Melchior , J.-L. Battaglia , A. Oustaloup (2014). Nonlinear thermal system identication using fractional Volterra series. Contr. Eng. Pract. , vol. 29, pp. 5060. D. Wang , X. Wang , P. Han (2010). Identification of thermal process using fractional order transfer function based on intelligent optimization. In: Proceedings of 2010 IEEE/ASME International Conference on Mechatronic and Embedded Systems and Applications, Qingdao, pp. 498503. Y. Huang , F. Guo , Y. Li , Y. Liu (2015). Particle estimation of fractional-order chaotic systems by using quantum parallel particle swarm optimization algorithm. PLoS ONE , vol. 10(1), pp. e0114910. DOI: 10.1371/journal.pone.0114910. W. Du , Q. Miao , L. Tong , Y. Tang (2017). Identification of fractional order system with unknown initial values and structure. Phys. Lett. A , vol. 381(23), pp. 19431949. M. Aoun , R. Malti , O. Cois , A. Oustaloup (2002). System identification using fractional Hammerstein models. IFAC Proc. Vol. , vol. 35(1), pp. 265269. D. Maiti , M. Chakraborty , A. Konar (2008). A novel approach for complete identication of dynamic fractional order systems using stochastic optimization algorithms and fractional calculus. In: Proceedings of IEEE international conference on Electrical and Computer Engineering (ICECE 2008) , Dhaka, Bangladesh, pp. 867872. S. Victor , R. Malti , P. Melchior , A. Oustaloup (2011). Instrumental variable identication of hybrid fractional Box-Jenkins models. IFAC Proc. Vol. , vol. 44(1), pp. 43144319. D.-Y. Liu , T.-M. Laleg-Kirati , O. Gibaru , W. Perruquetti (2013). Identication of fractional order systems using modulating functions method. In: Proceedings of 2013 American Control Conference (ACC) , Washington, DC, United States, pp. 16791684. O. Cois , A. Oustaloup , T. Poinot , J.-L. Battaglia (2001). Fractional state variable filter for system identification by fractional model. In: Proceedings of 2001 European Control Conference (ECC) , Porto, pp. 24812486. D.V. Ivanov , A.V. Ivanov (2017). Identification fractional linear dynamic systems with fractional errors-in-variables. J. Phys. Conf. Ser. , vol. 803(1), pp. pages 012057012057012058. M. Aoun , R. Malti , F. Levron , A. Oustaloup (2007). Synthesis of fractional Laguerre basis for system approximation. Automatica , vol. 43, pp. 16401647. Y. Li , X. Meng , B. Zheng , Y. Ding (2015). Parameter identication of fractional order linear system based on Haar wavelet operational matrix. ISA T. , vol. 59, pp. 7984. Y. Tang , H. Liu , W. Wang , Q. Lian , X. Guan (2015). Parameter identification of fractional order systems using block pulse functions. Signal Process. , vol. 107, pp. 272281. K.B. Datta , B.M. Mohan (1995). Orthogonal Functions in Systems and Csontrol . Singapore: World Scientific Publishing Co. Pte Ltd.
Design of Fractional Order Controllers using Triangular Strip Operational Matrices M. Manfred , Z. Evanghelos (1989). Robust Process Control . Englewood Cliffs, NJ: Prentice Hall. H.W. Bode (1945). Network Analysis and Feedback Amplier Design . New York: Van Nostrand. A. Oustaloup (1991). La Commade CRONE: Commade Robuste dOrdre Non Entier . Paris: Hermes. I. Podlubny (1999). Fractional-order systems and PID controllers. IEEE T. Automat. Contr. , vol. 44, pp. 208214. C. Zhao , D. Xue , Y.Q. Chen (2005). A fractional order PID tuning algorithm for a class of fractional order plants. In: Proceedings of the IEEE International Conference on Mechatronics & Automation , Niagara Falls, Canada, pp. 216221. I. Petras (2012). Tuning and implementation methods for fractional-order controllers. Fract. Calc. Appl. Anal. , vol. 15(2), pp. 282303. R. Caponetto , G. Maione , A. Pisano , M.R. Rapaic , E. Usai (2013). Analysis and shaping of the self-sustained oscillations in relay controlled fractional-order systems. Fract. Calc. Appl.
Anal. , vol. 16(1), pp. 93108. V. Feliu-Batlle , R. Rivas-Perez , L. Sanchez-Rodriguez , M.A. Ruiz-Torija (2009). Robust fractional order PI controller implemented on a laboratory hydraulic canal. J. Hydraul. Eng. , vol. 135, pp. 271282. Y. Luo , Y.Q. Chen (2009). Fractional-order proportional derivative controller for robust motion control: Tuning procedure and validation. In Proc. of American Control Conference , Hyatt Regency Riverfront, St. Louis, MO, pp. 14121417. Y. Luo , Y.Q. Chen (2012). Stabilizing and robust fractional order PI controller synthesis for rst order plus time delay systems. Automatica , vol. 48, pp. 21592168. C.A. Monje , Y.Q. Chen , B.M. Vinagre , D. Xue , V. Feliu (2010). Fractional-order Systems and Controls: Fundamentals and Applications . London: Springer-Verlag. C.A. Monje , A.J. Calderon , B.M. Vinagre , Y. Chen , V. Feliu (2004). On fractional controllers: Some tuning rules for robustness to plant uncertainties. Nonlinear Dynam. , vol. 38, pp. 369381. M. Talebpour , Y.M. Roshan , S. Mohseni (2009). Developing robust FOPID controllers based on fuzzy set point weighting algorithm. Fract. Calc. Appl. Anal. , vol. 12, pp. 373390. M. Zamani , M. Karimi-Ghartemani , N. Sadati (2013). FOPID controller design for robust performance using particl swarm optimization. Fract. Calc. Appl. Anal. , vol. 10, pp. 169188. P. Lanusse , J. Sabatier (2011). PLC implementation of a CRONE controller. Fract. Calc. Appl. Anal. , vol. 14(4), pp. 505522. M.K. Bouafoura , N.B. Braiek (2010) PID controller design for integer and fractional plants using piecewise orthogonal functions. Commun. Nonlinear Sci. Numer. Simulat. , vol. 15, pp. 12671278. B. Liao , R. Luus (2005). Comparison of the Luus-Jaakola optimization procedure and the genetic algorithm. Eng. Optimiz. , vol. 37, pp. 381396. J. Robinson , Y. Rahmat-Samii (2004). Particle swarm optimization in electromagnetics. IEEE T. Antenn. Propag. , vol. 52, pp. 397408. 281 M. Tabatabaei , M. Haeri (2011). Design of fractional order proportional integral derivative controller based on moment matching and characteristic ratio assignment method. Proc. IMechE., Part I: J. Syst. Cont. Eng. , vol. 225, pp. 10401053. F. Merrikh-Bayat , M. Karimi-Ghartemani (2010). Method for designing PID stabilizers for minimum-phase fractional-order systems. IET Control Theory Appl. , vol. 4, pp. 6170. M.A. Sahib (2015). A novel optimal PID plus second order derivative controller for AVR system. Eng. Sci. Technol. , vol. 18, pp. 194206. Y. Tang , M. Cui , C. Hua , L. Li , Y. Yang (2012). Optimum design of fractional order PID controller for AVR system using chaotic ant swarm. Expert. Syst. Appl. , vol. 39, pp. 68876896. I. Podlubny (2000). Matrix approach to discrete fractional calculus. Fract. Calc. Appl. Anal. , vol. 3, pp. 359386.
Rational Integer Order System Approximation for Irrational Fractional Order Systems K. Biswas , S. Sen , P.K. Dutta (2006). Realization of a constant phase element and its performance study in a differentiator circuit. IEEE T. Circ. Syst. II , vol. 53(9), pp. 802806. I.S. Jesus , J.A.T. Machado (2009). Development of fractional order capacitors based on electrolyte processes. Nonlinear Dynamics , vol. 56(12), pp. 4555. A. Oustaloup , F. Levron , B. Mathieu , F.M. Nanot (2000). Frequency-band complex noninteger differentiator: Characterization and synthesis. IEEE T. Circ. Syst. I , vol. 47(1), pp.2539. D. Xue , C. Zhao , Y.Q. Chen (2006). A modified approximation method of fractional order system. In: Proceedings of the 2006 IEEE International Conference on Mechatronics and Automation , pp.10431048. D. Xue , Y. Chen (2005). Sub-optimum H2 rational approximations to fractional order linear systems. In: ASME. International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, Volume 6: 5th International Conference on Multibody Systems, Nonlinear Dynamics, and Control, Parts A, B, and C , pp. 15271536. G.E. Carlson , C.A. Halijak (1964). Approximation of a fractional capacitors (1/s)(1/n) by a regular Newton process. IEEE T Circ. Theory CT , vol. 11(2), pp. 210213. S. Das , S. Saha , A. Gupta , S. Das (2011). Analog realization of fractional order hybrid differentiators via Carlsons approach. In: Proceedings of 2011 International Conference on Multimedia, Signal Processing and Communication Technologies , pp. 6063. A. Charef , H.H. Sun , Y.Y. Tsao , B. Onaral (1992). Fractal system as represented by singularity function. IEEE T. Automat. Contr. , vol. 37(9), pp.14651470.
L. Meng , D. Xue (2012). A New Approximation Algorithm of Fractional Order System Models Based Optimization. J. Dyn. Syst. , vol. 134(4), pp. 044504-1044504-7. K. Matsuda , H. Fujii (1993). H-optimized wave-absorbing control: Analytical and experimental results. J. Guid. Control. Dynam. , vol. 16(6), pp.11461153. B.T. Krishna (2011). Studies on fractional order differentiators and integrators: A survey. Signal Processing , vol. 91, pp.386426. C.-H. Wang , C.-Y. Chen (2012). Finding the integer order systems for fractional order systems via fractional operational matrices. In: Proceedings of 2012 9th IEEE International Conference on Networking, Sensing and Control , Beijing, pp. 267270. D. Maiti , A. Konar (2008). Approximation of a fractional order system by an integer order model using particle swarm optimization technique. In: IEEE Sponsored Conference on Computational Intelligence, Control and Computer Vision in Robotics & Automation , India, pp. 149152. Z. Gao , X. Liao (2012). Rational approximation for fractional-order system by particle swarm optimization. Nonlinear Dyn. , vol. 67, pp.13871395. M. Khanra , J. Pal , K. Biswas (2010). Rational approximation of fractional order SISO System using squared-magnitude function. In: Mathematical Methods in Engineering International Symposium , Coimbra, Portugal. 309 M. Khanra , J. Pal , K. Biswas (2010). Rational approximation of fractional operator: A comparative study. In: 2010 International Conference on Power, Control and Embedded Systems (ICPCES) , Allahabad, pp. 15. O. Wing , Y.-L. Jiang , Q.-J. Yu (1998). Rational approximation of irrational functions by linear fractional transformations. IEEE T. Circ. Syst. I , vol. 45(11), pp. 12161221.
Numerical Method for Solving Fractional-Order Optimal Control Problems J.T. Betts (2001). Practical Methods for Optimal Control Using Nonlinear Programming . Philadelphia: SIAM. 328 J.T. Betts , S.O. Erb (2003). Optimal low thrust trajectories to the moon. SIAM J. Appl. Dyn. Syst. , vol. 2, pp.144170. A.B. Malinowska , D.F.M. Torres (2012). Introduction to the Fractional Calculus of Variations . London: Imperial College Press. R. Almedia , S. Pooseh , D.F.M. Torres (2015). Computational Methods in the Fractional Calculus of Variations . London: Imperial College Press. A.B. Malinowska , T. Odzijewicz , D.F.M. Torres (2015). Advanced Methods in the Fractional Calculus of Variations . New York: Springer Briefs in Applied Sciences and Technology. L.S. Pontryagin , V. Boltyanskii , R. Gamkrelidze , E. Mischenko (1962). The Mathematical Theory of Optimal Processes . New York: John Wiley & Sons. M.H. Heydari , M.R. Hooshmandasl , F.M. Maalek Ghaini , C. Cattani (2016). Wavelets method for solving fractional optimal control problems. Appl. Math. Comput. , vol. 286, pp.139154. E. Keshavarz , Y. Ordokhani , M. Razzaghi (2015). A numerical solution for fractional optimal control problems via Bernoulli polynomials. J. Vib. Cont. , vol. 22(18), pp.38893903. S. Nemati (2016). A spectral method based on the second kind Chebyshev polynomials for solving a class of fractional optimal control problems. Sahand Commun. Math. Anal. , vol. 4(1), pp.1527. A.H. Bhrawy , E.H. Doha , D. Baleanu , S.S. Ezz-Eldien , M.A. Abdelkawy (2015). An accurate numerical technique for solving fractional optimal control problems. Proc. Romanian Acad. , vol. 16(1), pp.4754. M.H. Heydari , M.R. Hooshmandasl , A. Shakiba , C. Cattani (2016). An efficient computational method based on the hat functions for solving fractional optimal control problems. Tbilisi Math. J. , vol. 9(1), pp.143157. A. Lotfi , S.A. Yousefi , M. Dehghanb (2013). Numerical solution of a class of fractional optimal control problems via the Legendre orthonormal basis combined with the operational matrix and the Gauss quadrature rule. J. Comput. Appl. Math. , vol. 250, pp.143160. S.A. Yousefi , A. Lotfi , M. Dehghan (2011). The use of a Legendre multiwavelet collocation method for solving the fractional optimal control problems. J. Vib. Cont. , vol. 17(13), pp.20592065. A. Lotfi , S.A. Yousefi (2014). Epsilon-Ritz method for solving a class of fractional constrained optimization problems. J. Optimiz. Theory. App. , vol. 163, pp.884899.
View publication stats